The Codex backend (client_version 0.156.0) served seven list+supported
models on 2026-09-22: gpt-6-astra, gpt-6-sol, gpt-6-luna, then
gpt-5.6-sol/terra/luna and gpt-5.5. Sol and Luna carry context 272000,
efforts low..max, default medium. POST /codex/responses with kigi's codex
headers answers 200 and completes for both.
`shift_enter_unavailable` kept its own list of brands and missed every
other reason the Kitty keyboard protocol gets skipped: JetBrains, screen,
tmux below 3.3 and tmux with `extended-keys off` all swallow it, yet the
prompt still advertised Shift+Enter there. It now reads
`kitty_skip_reason` directly. Two skips deliver the modifier another way
and stay available: Apple Terminal, where `route_enter` polls
CoreGraphics, and Windows Terminal, where the console key record carries
SHIFT — a bare ConHost, only optimistically refined to Windows Terminal,
keeps advertising Alt+Enter.
The old list also carried a `VTE_VERSION >= 8200` cutoff, citing VTE
0.82.0 as the first release with the protocol. It is not: at tag 0.82.0
the string "kitty" appears nowhere in `src/parser-seq.py`, `src/vte.cc`
or `src/widget.cc`, and `_vte_keymap_GDK_Return` encodes only Alt
(`ESC CR`), so Shift+Enter leaves a VTE terminal as the same bare CR as
Enter. Every VTE now advertises Alt+Enter, and `/terminal-setup` stops
telling GNOME Terminal users to upgrade to a version that would not
help. It also stops advertising Alt+Enter to WezTerm, which binds that
chord to ToggleFullScreen.
The Codex backend served five `visibility=list` + `supported_in_api`
models on 2026-09-21: `gpt-6-astra` ahead of `gpt-5.6-sol/terra/luna`
and `gpt-5.5`. Astra carries context 272000 and efforts
low/medium/high/xhigh/max, default medium; the `ultra` tier it also
advertises has no kigi level and is left out. `gpt-reserve` and
`codex-auto-review` stay hidden.
Inference checked against the live backend: `POST
/backend-api/codex/responses` with `model: gpt-6-astra` and the codex
`User-Agent` kigi sends answers 200 and streams, so the model works on
the compiled-in path.
Reverts c61977c. The live listing needed no codex CLI on the user's
machine — kigi issued the request itself and `client_version` was a
compiled-in string — but the project keeps the catalog in the binary, so
the fetch path, the `OpenaiCodex` listing dialect, its parser and the
pinned client version are gone again, and the codex `User-Agent` returns
to its previous value.
The models.dev provider rename `kimi-for-coding` → `kimi-code-plan-cn`
rode in the same commit and stays: the regenerated snapshot ships under
the new id.
`xai` and `openai` carry `restrict_to_enriched`, so a live listing id the
enrichment catalog does not know is dropped. The bundled snapshot knew
neither `grok-4.7` nor `gpt-6-astra`, which left both invisible offline
and until the 24h models.dev refresh landed.
Regeneration also tripped the generator's fail-fast: models.dev renamed
the Kimi coding provider. `kimi-for-coding` is now `kimi-code-plan-cn`
(api.kimi.com — kigi's endpoint), with `kimi-code-plan-global` the
api.kimi.ai twin; the model set is unchanged. Both registry rows, the
generator target and the two snapshot lookups follow the new id, and the
cache `keep_set` guard invalidates any cache filtered under the old one.
The ChatGPT Codex catalog was compiled in: four models read once from the
Codex CLI cache. The backend moved on — `gpt-6-astra` shipped, and
`gpt-5.3-codex-spark` left — so the picker showed a table that no longer
matched the account.
`GET https://chatgpt.com/backend-api/codex/models?client_version=<v>`
serves that table. Verified against a live ChatGPT account: the OAuth
Bearer alone returns 200, the query is required (400 without it), and the
backend gates the catalog on it — `0.104.0` answers with
`codex-auto-review` alone, `0.150.0` omits `gpt-6-astra`, `0.155.1`
serves the full set. `CODEX_CLIENT_VERSION` pins the current upstream
release (openai/codex rust-v0.155.1) and the codex `User-Agent` now
derives from it, so the two can never drift.
`ListingDialect::OpenaiCodex` + `parse_openai_codex_listing` keep the
`visibility == "list" && supported_in_api` models and carry the wire
context window and per-model reasoning levels; `hardcoded_catalog` and
its short-circuit in the fetch path are gone. The catalog advertises an
`ultra` tier kigi has no level for — `think_efforts_to_options` drops it
with a warning, as it does any unknown token.
`cargo about generate about.hbs -o THIRD-PARTY-NOTICES.md` (cargo-about
0.9.2) on the 0.1.16 lock: kigi crates at 0.1.16, rustls 0.23.45,
rustls-webpki 0.103.15, aws-lc-rs 1.18.1, aws-lc-sys 0.45.0. The set of
listed crates is otherwise unchanged; the large diff is block order.
`e2e_default_model_with_session_routes_to_proxy` and
`e2e_duplicate_model_field_both_entries_survive` assert the default
model's base URL. They were not serial and read `KIGI_CODE_BASE_URL`
through `EndpointsConfig`, while two serial tests (auth_method.rs,
models_fetch.rs) point that variable at wiremock. When the runs
overlapped, the base URL came back as the wiremock URI: the Linux suite
failed in 2 of 13 full runs. Presetting the variable fails both tests
every time, on v0.1.15 as well.
Both tests are now `#[serial]` and unset the variable for their scope,
the same cure `e2e_enterprise_endpoints_only_no_model_override` got for
the moonshot variable. With the variable preset they pass.
The sync record now ends at upstream `a28ee2b` (2026-09-17, ≈v1.0.35).
The items evaluated in this window and left for a later cycle are listed
with the reason each one waits, so the next sync does not triage them
again. `pager.toml` joins that list: its writer still ends in a bare
`fs::rename`.
Two stale lines go: `ultra` left the reasoning-effort vocabulary in
649c5f5, yet the file still described it as a codex-only tier.
`cargo deny check advisories` failed on rustls 0.23.42: TLS 1.3 handshake
messages were accepted across encryption-level boundaries. `cargo update
-p rustls` moves rustls to 0.23.45 and, with it, rustls-webpki to
0.103.15, aws-lc-rs to 1.18.1 and aws-lc-sys to 0.45.0.
The same update also re-picked six unrelated edges among versions
already in the lock (errno, quinn-udp, rustix, rustls-platform-verifier
and tempfile from windows-sys 0.52.0 to 0.59.0; tempfile from getrandom
0.4.3 to 0.3.4). Those edges are restored, so the diff is the four
packages only: version and checksum, sixteen lines. `cargo metadata
--locked` accepts the lock.
Advisories pass. `cargo clippy --workspace --all-targets --locked` is
clean on Windows.
The osascript fallback hands the pasted raster to kigi through a file.
Every kigi process used the same three paths in the shared temp dir
(`kigi-clipboard-probe.{png,tiff,jpg}`). Two sessions pasting at once, or
a stale file from an earlier probe, could swap the file between the
AppleScript's write and kigi's read: the wrong image went to the model.
`ProbeTemps` owns a fresh 0700 directory per probe and removes it on
drop, so the hand-written removals are gone. The script builder, the
raster reader, and the read flow move out of the cfg(macos) module into
`clipboard/osascript_probe.rs`, which compiles under test on every
platform; the flow takes the script runner as a closure. The script text
with a scratch dir is byte-identical to the old builder's (checked once
against the old function, then the copy was dropped).
A file paste needs no scratch file. When the directory cannot be made,
the probe warns, sends the file-URL half of the script only, and answers
file URLs as before; with no file URL either, the scratch error surfaces
instead of an empty board. `get_image` needs the file and fails fast.
Tests (all platforms): probes never share a path, the directory goes
away with the probe, mode 0700 on unix (Linux under WSL), the reported
class is read from this probe's file, file URLs survive a missing
scratch dir, and no dir plus no file URL is an error. The remaining
cfg(macos) callers were type-checked in a stubbed scratch crate; the
warm-cache workflow builds both darwin targets before the tag.
Ported from upstream (grok-build shared/clipboard.rs ProbeTemps and
osascript_attachments, 1.0.33).
Every config.toml writer wrote a tmp beside the path and renamed it over
the path. With config.toml symlinked into a dotfiles repo, the rename
replaced the link with a regular file: the repo copy went stale and the
link was gone.
`config_write_dest` resolves the destination before a writer names its
tmp. For the user's own config path a symlink resolves to its referent,
so the tmp sits on the referent's filesystem and the rename stays atomic.
A dangling link is an error, not a clobber. All eight writers call it:
`save_config`, the four MCP writers, and the three import writers.
Only the user's config path is followed, and only while
`user_kigi_home()` resolves. `kigi mcp add --scope project` and the
Claude project import write `.kigi/config.toml` inside the repo; a
hostile repo could ship that file as a symlink to any file the user
owns. With no home, `kigi_home()` falls back to `./.kigi`, which is that
same project tree, so nothing is followed then. Untrusted paths come
back unchanged and the link itself is replaced, as before. AGENTS.md
records the rule.
Tests (unix; run on Linux under WSL): helper cases including the no-home
case, a project-scope save that spares the link's target, and an
end-to-end save under KIGI_SHARE_DIR. The end-to-end test fails with
following disabled. Windows symlinks were not exercised: the host lacks
the privilege.
Ported from upstream (grok-build config/fs_atomic.rs LeafSymlink::Follow,
1.0.35): the rule, not the 1375-line module.
`read_file` exempted SKILL.md from every limit, and both injection paths
(`/skill` expansion and the `skill` tool) pasted the whole body. A 220 KB
skill put 55k tokens into the context on each call.
`cap_skill_body` cuts an over-cap body at the last line boundary that
fits (a char boundary when the body is one long line) and appends a note
that points at offset and limit. It runs before `apply_substitutions`, so
the arguments land after the cut. Both injection paths call it and log
the cut.
`read_file` returns a SKILL.md whole only while it fits the cap. Over the
cap it reads like any other file: offset and limit apply, and a window
that is still too large answers FileTooLarge. Without that the note would
send the model to a read that ignores its range.
`skill_file_skips_token_limit` pinned the exemption; it gives way to
`skill_file_over_cap_returns_file_too_large` and
`skill_file_over_cap_honors_offset_and_limit`, as upstream did.
Ported from upstream (grok-build tools skills/skill.rs cap_skill_body and
read_file skill_full, 1.0.33). Upstream's `max_output_bytes` read budget
in the same hunk is a feature and stays out.
A soft-wrapped row and the next row share a byte: the row's exclusive end
is the next row's start. Up or Down with a goal column past the target
row's width put the cursor on that byte, so it drew on the wrong row. Up
made no progress and Down skipped a row.
`row_with_end` tells a soft wrap from a hard break (byte-adjacent rows
are soft). `move_to_display_col_on_line` takes the row and its `RowEnd`
and clamps a soft row to its last grapheme, or to the start of an element
that closes the row. A hard-break row still reaches its newline.
`clamp_to_line` now cuts on graphemes, not chars, so a ZWJ cluster stays
whole.
Kigi differs from upstream's premise that elements wrap whole: with no
display override in the buffer, `wrapped_lines` uses plain textwrap and
an element can span rows. Snapping to such an element's start would
leave the target row (Down from byte 0 stuck at 0), so an element that
began on an earlier row keeps the boundary `display_col_to_buffer_pos`
chose, as before this change.
`wrapped_navigation_with_zwj_graphemes` asserted the bug: Up from the
end stayed on the last row (byte 22). It now expects row 1's cluster
start (byte 11), as upstream does.
Tests: upstream's four cases, the row-spanning element, and a seeded
property test (500 cases): each Up or Down moves exactly one row. The
property test fails with the clamp disabled.
Ported from upstream (grok-build ratatui-textarea/textarea.rs, 1.0.28).
`try_call_tool` wrapped `call_tool` in a bare `tokio::time::timeout`.
On a timeout or a cancelled turn the future dropped and the server kept
working: no `notifications/cancelled` ever left the client.
`call_tool_cancel_aware` sends the request through
`send_cancellable_request` and holds a guard armed with the request id.
A settled call disarms it. A dropped future (cancelled turn) spawns the
cancel from the guard's drop. A timeout sends the cancel inline, before
returning: the HTTP timeout arm resets the transport next, and dropping
the last service handle killed a spawned cancel before it left.
The inline send waits 500 ms at most. rmcp 2.2.0 awaits its own
timeout-cancel without a bound, and its streamable-HTTP worker awaits
each POST in turn, so against a JSON-mode server the cancel queues
behind the hung call: with rmcp's timeout path the call returned after
3.0 s against a 1 s budget. That server class cannot receive the cancel
until rmcp 3.x; the budget elapses, a warn is logged, and AGENTS.md
records the limitation. stdio and SSE-mode servers receive it.
Tests: dropped call, timed-out call followed by a service drop, settled
call; the HTTP outer-timeout test now asserts the return time.
Ported from upstream (grok-build mcp/servers.rs call_tool_cancel_aware,
1.0.33); upstream leans on rmcp 3.2's timeout path.
`reasoning_effort = "High"` in config.toml failed the whole parse: the
derived Deserialize matched lowercase variants only, while FromStr
lowercased its input. Deserialize now routes through FromStr, so both
paths share one vocabulary. Unknown tokens and non-strings still error.
Ported from upstream (grok-build sampling-types/types.rs, 1.0.35).
Sync record now ends at upstream `3794978` (2026-09-09, ≈v1.0.24). Lists
the two limitations the sync left: the Windows frame writer still prints
through std's stderr lock, and the bwrap builder re-stats /data.
Claude-Session: https://claude.ai/code/session_01RUqV8HVGXcAETyYbZ2tPQ4
`extract_changed_files` parses `diff --git a/<old> b/<new>` headers.
A host `diff.mnemonicprefix=true` or `diff.noprefix=true` rewrote them
(`c/ w/`, or none), so every tracked change vanished from `changed_files`
and the two evidence git-diff tests failed on such hosts. Both capture
paths now pass `--src-prefix=a/ --dst-prefix=b/`, which override both
config keys. Regression test sets each key in the temp repo.
Claude-Session: https://claude.ai/code/session_01RUqV8HVGXcAETyYbZ2tPQ4
A devbox-based profile on a host with /data warned and fell back to
Landlock when bubblewrap was missing or failed to exec, and Landlock
grants the workspace under /data. requires_data_write_deny (devbox
inheritance and an existing /data; a stat error other than NotFound
counts as present) now joins requires_read_deny in the fail-closed
startup decision, and the /data ro-bind is added only under that
predicate, so a missing /data no longer makes bwrap fail to exec.
Ported from upstream (grok-build sandbox/lib.rs requires_data_write_deny).
Claude-Session: https://claude.ai/code/session_01RUqV8HVGXcAETyYbZ2tPQ4
A project-scoped agent definition from an untrusted repository could
declare inline stdio mcpServers and the child session started that
command, while inline hooks on the same definition were already gated on
folder trust. The mcpServers list now uses the same gate.
By-name agent lookup joined the requested name onto each searched
directory, so a name carrying a separator or ".." resolved a file outside
that directory under its trust scope; both lookup entry points now reject
any name that is not a single path component.
Ported from upstream (grok-build subagent/handle_request.rs mcpServers
gate), extended with the name guard.
Claude-Session: https://claude.ai/code/session_01RUqV8HVGXcAETyYbZ2tPQ4
The pager rendered "Compaction failed." with the error dropped, and the
shell's unsuppressed auto-compaction failure sent an empty error. Both
now carry the decoded first line of the failure (retry guidance plus the
detail), decoded from the ACP data rather than its JSON-escaped Display,
propagated unchanged through the compact extension, capped for a banner.
Endpoint credentials never reach the banner: the sampler redacts the
parsed endpoint (userinfo, sensitive query values, fragment via the
repository sanitizer) in every 401 message and the appended request URL,
and the shell redacts every URL in the error text before line selection,
including URLs whose password holds whitespace or that follow a
multi-byte delimiter.
Ported from upstream 1.0.13 (grok-build dispatch/prompt.rs,
session/compaction.rs), extended with the redaction.
Claude-Session: https://claude.ai/code/session_01RUqV8HVGXcAETyYbZ2tPQ4
With fd 2 dead (a closed pane, an exited pipe reader) every write fails,
eprintln! panics on that failure, and under panic = "abort" the process
dies by SIGABRT with a crash report on the next launch instead of its
exit code. Exit reports, headless plain-format diagnostics, update-exit,
setup-result and leader-management messages now write through
best_effort_stderr::eprint_line: one attempt, failure reported, never a
panic.
Ported from upstream 1.0.19 (grok-build best_effort_stderr.rs).
Claude-Session: https://claude.ai/code/session_01RUqV8HVGXcAETyYbZ2tPQ4
str::lines() splits on neither separator, so a paste carrying them counted
as one line and the invisible characters broke the textarea's width model.
Bare \r, U+2028 and U+2029 now normalize to \n on every paste route (the
prompt, the inline paste key and the previous-message editor); dropped
file paths keep their bytes verbatim because a separator in a filename is
payload.
Ported from upstream 1.0.18 (grok-build prompt_widget normalize_line_breaks).
Claude-Session: https://claude.ai/code/session_01RUqV8HVGXcAETyYbZ2tPQ4
An attachment over the side or area caps whose re-encode did not shrink
the byte count was returned unchanged at its original dimensions, so the
provider still received an over-cap image. The byte-saving short-circuit
now applies only within the caps; a resize that would fall under the
vision minimums keeps the original on the existing fallback path.
Ported from upstream 1.0.13 (grok-build session/image_normalize.rs).
Claude-Session: https://claude.ai/code/session_01RUqV8HVGXcAETyYbZ2tPQ4
AutoCompactStarted carries the count the trigger fired on, but only
AutoCompactCompleted refreshed the context bar, so the "N% full" banner
sat next to a stale pre-turn number until the next stream update. Both
lifecycle updates now refresh the bar, on the root view and on subagent
views.
Ported from upstream 1.0.12 (grok-build acp_handler/session_notification.rs).
Claude-Session: https://claude.ai/code/session_01RUqV8HVGXcAETyYbZ2tPQ4
The queued /compact command drained into Effect::Compact with only the
session id, so the instructions typed after the command never reached the
shell's compaction prompt (which already accepts userContext). The effect
carries the parsed instructions, trimmed so an edited queue row with
leading whitespace still counts.
Ported from upstream 1.0.22 (grok-build dispatch/queue.rs).
Claude-Session: https://claude.ai/code/session_01RUqV8HVGXcAETyYbZ2tPQ4
A terminal that stopped draining the pty parked the writer thread inside
its tty write while it held the stderr lock; the next inline stderr write
from the event loop (mouse toggle, notification, kitty clear) deadlocked
the loop, and /quit hung on the unbounded writer join and the teardown
writes.
EscapeWriter queues out-of-band escapes on the frame writer: the mouse
capture toggle, notifications (title, progress, protocol), the auth
cursor, gboom keyboard flags, the wrap clipboard request, and the kitty
clears (queued per agent view and merged into the frame's post-flush,
nested subagent views included). The $EDITOR/$PAGER handoff aborts when
the writer does not drain. WriterThread::join_within and the teardown
run under a 2 s grace; after a stall fd 2 is silenced, the exit hints,
relaunch and update prints are skipped, and the panic hook keeps only its
log report.
Ported from upstream 1.0.19 (grok-build EscapeWriter, WriterJoin,
run_bounded_teardown), extended to the remaining inline writes.
Claude-Session: https://claude.ai/code/session_01RUqV8HVGXcAETyYbZ2tPQ4
A completed child parked forever behind a parent that never serviced its
command channel, or behind its own frozen chat-state actor, leaking the
child session thread, fs watchers and fds. Every await in the completion
path that rides another actor is now bounded: child chat-state reads
degrade to their fallbacks after 10 s, the usage-fold and not-applied acks
give up after 30 s (an unacked mark falls back to the coordinator's sticky
mark), list_tasks and reparent_notifications are bounded, and the
post-Shutdown worktree snapshot is bounded while removal is awaited in
full so no detached deletion can race a resume.
Ported from upstream 1.0.14 (grok-build subagent/handle_request.rs
PARENT_ACK_TIMEOUT / child_actor_query), reshaped without the drain-budget
const assert.
Claude-Session: https://claude.ai/code/session_01RUqV8HVGXcAETyYbZ2tPQ4
A Reasoning item counted its text and its encrypted blob as two separate
payloads, inflating the estimate for retained reasoning; the estimate now
takes the larger of the two (base64 runs 4/3 over the raw bytes).
Only compaction carried the provider-confirmed ratio into the reseeded
total; a rewind or a non-compaction replace sprang the context meter back
to the raw estimate. reseed_total_tokens applies the ratio (capped at the
pre-rewrite total) to every rewrite, and the truncate path uses it too.
Ported from upstream 1.0.1x (grok-build xai-chat-state actor/state.rs,
mutations.rs, queries.rs).
Claude-Session: https://claude.ai/code/session_01RUqV8HVGXcAETyYbZ2tPQ4
The slug-match step treated an entry whose api_backend equals the default
as unset and overwrote it with a same-slug sibling's backend, so
`api_backend = "chat_completions"` silently became messages/responses.
Config keys whose override sets api_backend are recorded and skipped by
the donor step; an unset backend still inherits.
Ported from upstream 1.0.22 (grok-build agent/config.rs
explicit_api_backend_keys).
Claude-Session: https://claude.ai/code/session_01RUqV8HVGXcAETyYbZ2tPQ4
The Messages stream failed every max_tokens stop with MaxTokensTruncation,
so a response that had already emitted complete tool calls before the
output budget ran out lost those calls and failed the turn. The stream
now completes with StopReason::ToolCalls when every collected call's
arguments parse (or are empty, the zero-argument convention); a cut call
or a text-only truncation still fails so compact-on-error keeps working.
The ChatCompletions and Responses streams already behaved this way.
Ported from upstream 1.0.1x LengthPolicy::CompleteToolCalls, reshaped
into the stream layer because Kigi has no caller that chooses a policy.
Claude-Session: https://claude.ai/code/session_01RUqV8HVGXcAETyYbZ2tPQ4
The persistence actor acked FlushAndAck unconditionally after writing its
pending merged notification: nothing was fsynced and every storage failure
was only warn-logged, so persist_ack, the reconnect flush and history
repair treated lost writes as persisted.
The barrier now writes the pending notification, reports the first storage
failure latched since the last barrier (every actor arm latches), and
fsyncs the whole session directory tree (files, directories, deletions).
Session directories are created durably (init and fork copy), and the
shared summary writer syncs its staged bytes before publishing and the
parent afterwards, so a concurrent /rename cannot publish unsynced bytes
over a synced summary.json. FlushComplete and the persist_ack, history
repair and end-of-turn callers propagate the io::Result; CopyFile fails on
a sync error but keeps the latch for the barrier.
Ported from upstream 1.0.1x (grok-build persistence.rs pending_write_error
+ durable storage helpers), reshaped without the dirty-file set.
Claude-Session: https://claude.ai/code/session_01RUqV8HVGXcAETyYbZ2tPQ4
Edit/Read/Grep rules matched only the raw argument string, so a workspace
symlink (or a quoted, `~`, `./`, `..`, or absolute spelling) reached a
denied file. The policy now resolves the path exactly as the tools do
(sanitize, expand `~`, display-cwd remap, join to the requesting session's
cwd), follows symlinks parent-first, and re-checks deny/ask rules on the
opened path and its target as absolute, `./`-relative and relative
spellings. Escalation only: no allow is granted through a link. A link
that cannot be followed forces a prompt that survives YOLO, auto mode and
the session edit grant, with event reason `unresolved_link`.
Permission requests now carry the requesting tool's cwd and display cwd
so a child worktree session sharing its parent's handle is checked where
it runs.
Ported from upstream 1.0.14 (grok-build permission/policy.rs symlink
re-check), reshaped for Kigi's lexical rule matcher.
Claude-Session: https://claude.ai/code/session_01RUqV8HVGXcAETyYbZ2tPQ4
Port of upstream grok-build 3794978 (snap6 "Reject ripgrep --hostname-bin
on always-safe and Auto routine paths"), widened where review showed the
same hole one layer down.
- `rg --pre` / `--hostname-bin` execute a program per file. The safe
list refused only `--pre`, and the Auto-mode routine classifier had
no rg flag check at all. One shared `rg_has_unsafe_flag` now refuses
both flags (bare or `=VALUE`) on a lowercased-basename head, scanning
every argument: rg's grammar (`--replace -- --pre`, `-ne --`) makes a
partial arity model bypassable, so a literal search for the flag name
costs a prompt instead of the auto-allow.
- The shell splitter handed allowlists undecoded source: a concatenation
(`--host""name-bin`, `''--hostname-bin`, `--p're'`) kept its quotes,
and double-quoted escapes (`"--host\<newline>name-bin"`) survived.
`PlainCommand` words are now the decoded argv (bash double-quote
rules, whole-node text minus the quotes so literal newlines and the
whitespace tree-sitter folds into the closing quote survive), with a
per-word `expandable` flag for unquoted glob/brace/tilde/backslash
syntax the shell would still expand (`--{pre,pre}`, a glob that can
match a file named `--pre`); the rg guard treats those as unsafe.
- Every prefix-based decision matches per argv element through one
matcher (built-in safe list, always-safe list, dangerous list, Auto
routine prefixes): a decoded word with inner whitespace (`rg `,
`rg /../payload`, one word spelled `git status`) can no longer pose
as a prefix. Saved grants match token for word (a grant saved from a
spaced argument prompts again). Saved denies keep the pre-0.1.15
check verbatim on a `legacy_words` spelling (raw content and raw
concatenations, wrappers peeled independently, space-boundary prefix)
so no deny narrows on upgrade.
Claude-Session: https://claude.ai/code/session_01RUqV8HVGXcAETyYbZ2tPQ4
Port of upstream grok-build 7581004 (1.0.22 "Auto mode no longer
instantly runs destructive git checkout -- commands"), routine_git.rs.
`ROUTINE_PREFIXES` auto-allowed `git checkout`, `git switch` and
`git stash` wholesale, so Auto mode ran `git checkout -- <file>` (a
working-tree discard) and `git stash drop`/`clear` without review.
Those verbs now go through a fail-closed operand classifier: `switch`
is routine with one ref operand and benign flags; `stash` only for
push/save/pop/apply/list/show/branch; `checkout` additionally needs a
branch-mode flag (`-b`, `-B`, `--orphan`, `--detach`, `-t`) anywhere in
the argument list before an operand counts, because a bare
`git checkout <name>` — `-` included — restores a tracked path when no
such ref resolves (a deliberate deviation from upstream, which allowed
`git checkout main` outright). Path-like operands, pathspec flags and
`--` fail closed. What is not routine goes to the model side-query, not
to a hard block.
Claude-Session: https://claude.ai/code/session_01RUqV8HVGXcAETyYbZ2tPQ4
Port of upstream grok-build 7581004 (1.0.22 window) folder_trust.rs.
`collect_repo_config_kinds` gated `.kigi/hooks` with `is_dir()` and
`.cursor/hooks.json` with `is_file()`, and `.claude/settings*.json` via
the settings-file scan, while the hook loader (`path_to_source`) takes
a directory at any of the four project hook paths as a hook source and
parses its JSON children. A clone carrying `.cursor/hooks.json/evil.json`
or `.claude/settings.json/evil.json` and nothing else resolved as
trusted through the no-config branch and ran its SessionStart hook.
The gate now treats any present path (file, directory, dangling
symlink) or an unreadable one at the four sources
`discover_hook_source_paths` loads from the git root as a hit.
Claude-Session: https://claude.ai/code/session_01RUqV8HVGXcAETyYbZ2tPQ4
Port of upstream grok-build 1.0.14 ("--sandbox strict now restricts
writes to ~/.grok/sessions only"), shaped for Kigi.
The strict profile granted the whole ~/.kigi read-write, so a sandboxed
command could plant a global SessionStart hook or flip
`[sandbox].profile` in config.toml for the next launch. Landlock cannot
carve a child out of a write grant, so the home is now read-only
except an explicit allowlist of state subtrees (run, sessions,
worktrees, worktree_pool, tasks, memory, logs, debug, crash, memtrace,
trace-exports); config.toml, auth.json, sandbox.toml, hooks/, bin/ and
vendor/ stay read-only. A symlinked state dir under the home is
followed only to the default ~/.kigi/sessions.
What the process itself writes there after enforcement moves or is
minted first:
- kigi-shell `apply_sandbox` mints the device id, extracts the bundled
rg and the bundled skills BEFORE enforcement for every profile that
restricts home writes (`restricts_home_writes`, custom profiles
extending strict included).
- The sandbox audit log lives at ~/.kigi/sessions/sandbox-events.jsonl.
- Leader lock/socket default to ~/.kigi/run/, the leader stderr log to
~/.kigi/logs/leader.log; discovery scans run/ and the legacy root,
and the default target prefers the current socket.
Root-level caches (version.json, agent_id, slash-mru.json, pager.toml,
claude_import_state.json) degrade best-effort under strict; documented.
Claude-Session: https://claude.ai/code/session_01RUqV8HVGXcAETyYbZ2tPQ4
Port of upstream grok-build 9684fa3 (1.0.11 window) `child_net.rs`.
The per-child seccomp filter matched seven socket syscalls by number and
allowed everything else, so a restricted child could still reach the
network through `sendmmsg` or an io_uring submission queue, and a
syscall issued through another arch table (x32 on x86_64) never matched
the deny list at all.
- Gate on the audit arch before matching `nr` and reject x32 numbers on
x86_64 (foreign-table syscalls now get EPERM instead of passing).
- Add `sendmmsg`, `io_uring_setup`, `io_uring_enter`, `io_uring_register`.
- Build the program once (`prebuilt_child_network_filter`) and hand the
slice to the `pre_exec` installer; the old code allocated the Vec after
`fork`, which is not async-signal-safe. All three callers updated.
- Tests evaluate the BPF program over synthetic `seccomp_data`.
Claude-Session: https://claude.ai/code/session_01RUqV8HVGXcAETyYbZ2tPQ4
After `/model`, the header kept the previous model's window ("15K /
1.0M" on a 262K model) until the next turn or a restart. The pager
caches the window inside its context snapshot (streaming `totalTokens`
updates refresh it from the catalog) and nothing refreshed the cache
when the session's model changed. Comparing cached numbers against
catalog windows cannot tell a catalog-derived total from a shell
effective limit (KIGI_DEBUG_CONTEXT_WINDOW, header upgrades), so the
shell, which installs the effective window, now reports it:
- kigi-shell: `handle_set_session_model` returns the installed window;
`model_switch::apply` stamps it as `totalContextTokens` in the
`session/set_model` response meta and carries it on the
`ModelChanged` broadcast (`context_window`, serde-default: old pagers
ignore it, old leaders omit it). The response meta previously
serialized the actor's inner `Result` and reported success on an
actor failure; that error now propagates.
- kigi-tui: `Effect::SwitchModel` reads the meta; `SwitchModelComplete`
carries `Ok(Option<u64>)`; the completion handler and the follower
`ModelChanged` arm call `AgentView::apply_context_window`, which
re-derives the snapshot (pct, free tokens) and reports whether it
changed so a follower redraws on a window-only change.
- Tests: the bare `/model <name>` dispatch path asserts the snapshot and
the drawn header; follower with/without the field; same-model new
window redraw; actor returns the window; serde round trip.
Claude-Session: https://claude.ai/code/session_01RUqV8HVGXcAETyYbZ2tPQ4
The GitHub account is restored, so github.com/ZacharyZhang-NY/Kigi-CLI is
the canonical repo again. Reverse the 2026-08 Gitea migration:
- .github/workflows/{release,warm-cache}.yml: restore the GitHub-hosted
pipeline dropped in 3e1652c, unchanged — hosted runners, rust-cache,
softprops release with every asset attached atomically.
- .gitea/workflows/release.yml: retire the Gitea edition (the runners are
being deregistered; git.zacharyzhang.com keeps the v0.1.12/v0.1.13
releases for old clients).
- kigi-env: update_base_url now defaults to the GitHub Releases API. The
updater already speaks GitHub wire format and sends a User-Agent; v0.1.14
is published to BOTH forges with identical assets so pre-0.1.14 installs
polling Gitea self-update onto the GitHub channel.
- install.sh / install.ps1: resolve releases from api.github.com.
- feedback, /docs, SECURITY.md, frontend worker + site: repo, issues and
raw-file links point at github.com / raw.githubusercontent.com.
- docs/RELEASE.md: GitHub edition of the checklist; Gitea runner sections
retired.
- version: 0.1.13 -> 0.1.14 (the cross-forge transition release).
Claude-Session: https://claude.ai/code/session_01A6RcwHhhsFYYrNdQB4ASuP
Second layer of platform gaps the mac/linux-only gate history left:
unix-only test imports/helpers now cfg-gated (mermaid, fast-worktree,
hunk-tracker, pager-render link test, three pty_e2e module decls),
clippy-suggested cleanups in Windows-cfg code (file-utils classifier,
tui console-mode restore, shared clipboard arboard-error flow, tools
terminal closure, workspace capability tail), and `cargo update` of h2
(RUSTSEC-2026-0258) and webbrowser (RUSTSEC-2026-0257) so `cargo deny
check advisories` passes again.
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01KSy9S31pZNWXtRZNVJ1sDf
The 0.1.13 sync had to reconstruct the fork point by content
fingerprinting; the sync record (baseline snapshot, last-evaluated
upstream commit, and where upstream hides its per-change lists) now
lives in AGENTS.md so the next sync starts from facts.
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01KSy9S31pZNWXtRZNVJ1sDf
The full-workspace gates had never run on this platform: collapsible
ifs in the Windows shell probe, a same-type cast in the crash handler,
an unused AsRawHandle import, and the sandbox smoke-test example
building its unix-only body (libc, kernel enforcement) unconditionally.
The example now stubs main off-unix; formatting reflow from cargo fmt.
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01KSy9S31pZNWXtRZNVJ1sDf
Port of upstream grok-build 1.0.0/1.0.1-window retry policy.
- is_retryable_api_status: 429 plus any 5xx except Cloudflare 525/526
(origin TLS handshake / invalid certificate — never clears on its
own), replacing the 500|502|503|504|520 allowlist so the 521-524,
529, and 530 edge pages an outage serves actually retry.
- The generic retry path clamps a server Retry-After to
MAX_RETRY_BACKOFF (30s) and jitters it: Cloudflare 52x pages ask for
60-120s, which across the budget stalled a turn ~28 minutes. The 429
path keeps the full (parse-capped) Retry-After under its low attempt
cap, and jitter is shared via jittered().
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01KSy9S31pZNWXtRZNVJ1sDf
Port of upstream grok-build 1.0.9. A bare positional http(s):// URL
with no transport flag was silently added as a stdio command whose
"command" was the URL. It now infers the HTTP transport with a notice;
extra args, --env, legacy --command/--url, and every explicit flag keep
their prior meaning, and scheme-less URL-looking commands still stay
stdio with the existing warning.
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01KSy9S31pZNWXtRZNVJ1sDf
Port of upstream grok-build 1.0.4/1.0.5. Two paths read $HOME only,
which is usually unset on Windows outside Git Bash:
- the global-gitignore fallback (prompt/ignore.rs) now uses
std::env::home_dir(), the crate-wide idiom;
- fast-worktree's resolve_kigi_home falls back to std::env::home_dir()
with the same dunce canonicalization, eliminating the documented
divergence from kigi_config::default_kigi_home. On Unix home_dir()
reads $HOME first, so behavior is unchanged where HOME is set.
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01KSy9S31pZNWXtRZNVJ1sDf
Port of upstream grok-build 0.2.113/0.2.114 fixes, extended to Kigi's
startup order.
- kigi-tty-utils runtime module: worker-cap policy (8 max) — tokio's
per-core default pins 100+ thread slots on many-core shared hosts
against RLIMIT_NPROC / pids.max, and later spawns die with EAGAIN.
- kigi-bin builds the main runtime with capped workers and reports an
out-of-threads host with a diagnostic + exit(1) instead of a panic.
- The pre-runtime optimization threads (HTTP warmup, early model
prefetch) spawn recoverably and degrade — their panic would have
preempted the runtime's own diagnostic.
- The per-session actor thread and its runtime report spawn/build
failure through the existing acp::Error / AgentBuildError boundaries
(new RuntimeBuild variant) instead of panicking the agent host.
- fsnotify's watcher thread spawn maps into FsNotifyError::WatcherStart
so callers degrade instead of crashing.
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01KSy9S31pZNWXtRZNVJ1sDf
Port of upstream grok-build 1.0.1 (landed upstream 2026-08-06). The
coordinator fire-and-forgot every spawn, so a wide fan-out (graph
waves, swarm bursts, model-driven Task storms) could exhaust file
descriptors.
- kigi-tools task/admission: SubagentLimits — 32 concurrent per session
by default (KIGI_MAX_CONCURRENT_SUBAGENTS), queue|fail behavior
(KIGI_SUBAGENT_LIMIT_BEHAVIOR), zero clamped to 1, read once at the
composition root; parse_positive_env helper.
- Each parent session gets one spawn semaphore (map entries live for
the process so a resumed session shares its predecessor's limit);
handle_subagent_request holds an owned permit for the whole child
run — queue awaits FIFO with a unified-log notice at saturation,
fail routes through send_pre_spawn_failure. After a queued wait, a
closed parent command channel (teardown) fails the spawn instead of
running against a dead parent.
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01KSy9S31pZNWXtRZNVJ1sDf
Port of upstream grok-build 1.0.1 (landed upstream 2026-08-06).
A child wedged in an uninterruptible kernel syscall (D-state) never
observes SIGKILL, and a background descendant can hold the output pipes
open past the kill; both previously stalled tool futures without bound.
- kigi-tty-utils: KILL_REAP_TIMEOUT (5s) + reap_killed_bounded — wait
for an already-killed child at most the bound, then abandon the corpse
to tokio's orphan reaper.
- local_terminal: pipe readers share their buffer with the spawner and
join under the reap bound on both exit arms (aborted readers are
awaited so no tail bytes race the take); the whole tree is enrolled in
a ProcessGroup and group-killed on timeout; the timeout arm returns
the synthetic result instead of waiting on the corpse. Grandchild
kill + prompt-return pinned by a ported unix regression test.
- kigi-tools: reap_killed_search_child wrapper; grep's four kill sites
reap bounded; glob's stderr goes to null (an unread pipe blocks rg on
warning backpressure) and its truncation kill reaps bounded.
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01KSy9S31pZNWXtRZNVJ1sDf
Port of upstream grok-build 0.2.105 + 0.2.118.
- CompactionToolChoice (auto default | none), resolved env
KIGI_COMPACTION_TOOL_CHOICE > [features].compaction_tool_choice >
remote settings. Some OpenAI-compatible servers reject
tool_choice:"none" with tools attached, failing every long-session
compaction; auto keeps them working (the summarization prompt already
forbids tool use). Threaded from agent + subagent spawn through
CompactionConfig into generate_session_compact; the Responses gate
counts hosted tools too, so a hosted-only request still carries the
configured choice.
- is_context_length_error recognizes the "current message ... exceeds
budget" phrasing, so that overflow is Fatal instead of burning the
retry budget on a deterministic failure.
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01KSy9S31pZNWXtRZNVJ1sDf
Port of upstream grok-build 0.2.103. session/prompt runs an async intake
preamble before enqueueing SessionCommand::Prompt on the actor mailbox; a
concurrent cancel could enqueue Cancel first, land on an idle actor as a
no-op, and the prompt then started a turn nothing would ever cancel.
The per-session prompt-intake lock becomes the dispatch lock: prompt()
holds it across intake as before, and cancel() now holds it around its
Cancel send, so a cancel orders behind the prompt it targets. Regression
test ported verbatim (unix-only harness, like its siblings).
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01KSy9S31pZNWXtRZNVJ1sDf
Port of upstream grok-build 0.2.113-window fixes:
- write_auth_json_atomic tmp path gains a monotonic sequence next to the
pid, so two concurrent in-process writers (background mint + proactive
refresher) cannot share one tmp path and tear each other's bytes; a tmp
write that fails partway is reclaimed instead of orphaned.
- New secure_file::ensure_owner_only_permissions (0o600 / Windows ACL)
tightens pre-existing world-readable credential files on read, backup,
write, replace, and restore. Best-effort with warns after durable
content — a chmod-only failure must not masquerade as a failed write —
and propagating in restore_prior_bytes.
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01KSy9S31pZNWXtRZNVJ1sDf
Gitea ignores `.github/workflows/` entirely once `.gitea/workflows/` is
non-empty, so these two files have had no effect since the release workflow
landed — and the GitHub repo they targeted is gone (404, suspended account),
so they cannot run there either.
They were not merely dead: warm-cache.yml's Monday cron kept firing on the
Gitea side before `.gitea/` existed, queueing jobs against GitHub-hosted
labels (`macos-14`, `windows-2022`, `ubuntu-24.04-arm`) that no self-hosted
runner advertises. Those sat queued forever, and its one dispatchable leg
failed on `rustup: command not found` — the same bare-image gap the Gitea
workflow now provisions around.
`.gitea/workflows/release.yml` is the only pipeline; docs/RELEASE.md already
documents it as such.
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
Every crate inherits `workspace.package.version`, so the single edit plus
the refreshed lockfile carries the release. The `v0.1.12` tag is what
triggers the build; release.yml gates the tag against this value, and
`cargo metadata --locked` passes so the build's `--locked` will too.
This is the first release cut on Gitea Actions. v0.1.6 through v0.1.9 were
all cancelled for want of a runner whose label matched anything, so nothing
has ever published from the new forge — `releases/latest` still 404s, which
is exactly why `kigi update` finds nothing today. Ships the link migration
in the previous commit, so the installers and self-updater published here
point at Gitea instead of a GitHub repo that no longer serves them.
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
The canonical repo is git.zacharyzhang.com now, so every URL a user can
actually reach had to move with it.
Self-updater: `PRODUCTION_ENDPOINTS.update_base_url` targets Gitea's
`/api/v1/repos/.../releases`, which serves the same wire format GitHub's
releases API did — `Release`/asset deserialisation is unchanged. The alpha
channel requests `?per_page=30&limit=30` because GitHub reads the page size
from `per_page` and Gitea from `limit`, and each ignores the other's, so one
URL stays correct on either.
Installers: same API base, and the documented one-liners become
`https://kigicli.dev/install.{sh,ps1}`. That domain's Worker proxies both
files from Gitea raw `main`, so the published install command survives any
future forge move without being reprinted everywhere. install.ps1 also drops
the GitHub-specific `Accept: application/vnd.github+json`.
Everything else user-facing: `/feedback` and the shell's feedback fallback
point at the Gitea issue tracker, `/docs` at the Gitea-rendered README, and
web_fetch's User-Agent at kigicli.dev. SECURITY.md now routes reports to the
issue tracker with a warning to withhold exploitable detail, because Gitea
has no equivalent of GitHub's private vulnerability reporting. Comments
naming `.github/workflows/release.yml` name `.gitea/` now.
Tests assert against these constants rather than literal URLs, so they
follow automatically; the one URL-shape assertion (production endpoints are
https) still holds.
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
`/prompt.txt` is the one-shot Mac Mini runner-setup prompt. It carries a
Gitea runner registration token once filled in, and a leaked registration
token lets anyone attach a runner to the account and receive our jobs —
including release jobs. Anchored with a leading slash so a legitimately
tracked prompt.txt deeper in the tree would still be versioned.
The .wrangler/ and node_modules/ lines were already sitting uncommitted in
the working tree for the untracked frontend/ worker; they belong in this
same file, so they ride along rather than being split out.
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
Gitea has no hosted runners, so every `runs-on` has to name a machine we
registered ourselves — and the labels in the first draft of this file did
not match what our machines advertise. The only online Linux runner
(`dokploy-runner-01`) offers `ubuntu-latest`/`ubuntu-24.04`/`ubuntu-22.04`,
never `linux-x86_64`, and Gitea's API cannot relabel a runner remotely, so
a `v*` tag would have queued forever against a label nothing served.
The three Linux jobs therefore run on `ubuntu-latest`. That runner is in
container mode against a bare runner-image, so those legs now provision
what they need per run: build-essential, pkg-config, unzip (protoc), tar
and jq, plus rustup when absent — installed with `--default-toolchain none`
so `rust-toolchain.toml` still decides the channel and `rustup show`
materialises it. `prepare` installs jq/curl the same way. macOS stays on
`macos-arm64` and Windows on `windows-x86_64`, both host mode with their
toolchains preinstalled, so the new provisioning step is gated on a
`linux:` matrix key rather than on `ext == 'tar.gz'`, which covers Darwin
too.
docs/RELEASE.md picks up the corrected label table, the fact that
`register --labels` is ignored once config.yaml supplies them, the
registration-token call being a POST (a GET parses as runner id 0 and
404s), and how each host-mode runner is supervised so it survives reboots
and crashes.
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
All 62 crates inherit `workspace.package.version`, so the single edit
plus the refreshed lockfile carries the release. The `v0.1.11` tag is
what triggers the build; release.yml gates the tag against this value.
Ships the sessionless-model-switch fix (f656627): `/model` from a
non-project launch directory now starts the session its deferred switch
drains into instead of dangling silently.
Launching kigi in a non-project directory (~/Downloads, ~, /tmp — anywhere
`is_project_dir` rejects) leaves the agent view session-less behind the
project-picker question, which only a PLAIN prompt can open. Slash
commands still execute there, so `/model <name> <effort>` stashed its
switch in `deferred_model_switch` — a stash that assumes a create is in
flight — and with none pending it dangled forever with zero feedback:
the picker rendered, the user chose model and effort, and nothing
changed. Not a 0.1.10 regression: 0.1.9 pty-reproduces identically; the
report correlated with the upgrade only because every earlier launch
happened to be from a project directory.
Apply the QueueCommand precedent ("queued slash work bypasses the
picker, so create the deferred session or it never drains") to both
deferral sites — Action::SwitchModel in the router and
set_default_model's no-session branch — via
skip_picker_and_create_session, whose in-flight guard already makes it a
no-op while a create is pending, so the racing-create case is unchanged.
SessionCreated then applies the stash through the existing
apply_deferred_model_switch path; pty-verified end-to-end from
~/Downloads (session created, model changed, effort applied).
All 62 crates inherit `workspace.package.version`, so the single edit
plus the refreshed lockfile carries the release. The `v0.1.10` tag is
what triggers the build; release.yml gates the tag against this value.
Also records the shared-SSRF-policy invariant in AGENTS.md, so a future
port does not reintroduce a second copy. The zero-egress entry is
compressed by the same number of lines to keep the file at 500.
The hook runner carried its own `is_blocked_ip`, a line-for-line copy of
the pre-hardening web_fetch predicate: loopback allowed unconditionally,
and no TEST-NET, 198.18/15, 240/4, 0.0.0.0/8, multicast, IPv6 site-local
or embedded-v4 wrapper coverage. Hook URLs come from settings, and
project settings from an untrusted repo are loaded today, so the gap is
reachable.
`kigi-hooks` already depends on `kigi-tools`, so the fix is to delete the
copy and call the shared predicate — no new crate, and no third
implementation to drift.
`allow_local` is on for hooks: a loopback hook receiver is a legitimate
local setup, which the old copy also allowed. It is now allowed only
when the URL names the host literally, so a public name resolving to
loopback is refused.
Range coverage now lives with the policy; the runner's tests pin what it
adds on top. The URL-scrubbing regression test moves off TEST-NET-1,
which the policy now blocks, onto a closed loopback port — faster, and
no real egress from a test.
`committed_edit_keeps_diff_line_backgrounds` passed alone and
single-threaded but failed under parallel execution, so the workspace
gate was red.
`terminal_native_lock_paints_only_native_colors` engages the
process-global terminal-native lock. While it is held `Theme::current()`
returns `terminal_default()`, whose `diff_*_bg` are all `Color::Reset`,
so no band is painted at all — and `current_kind()` reports a nominal
`KigiNight`, which is what made the failure read as a theme mismatch.
Two changes, both at the cause:
- hold the shared theme lock via `theme_cache::pin_theme()`, the helper
written for this and until now unused, so the lock cannot be engaged
mid-test
- assert the invariant (two distinct non-Reset bands) instead of exact
RGB: the entry line cache is keyed on the GLOBAL theme kind, so an
exact-color assertion races by construction. Span-level colors stay
pinned by the `tool::edit` tests that own them
The failure message now reports the theme, height, and the bands it
actually found, which is how the cause was located.
The hardcoded Codex catalog advertised `ultra` for gpt-5.6-sol and
gpt-5.6-terra, so selecting it sent `reasoning.effort = "ultra"` and the
Responses endpoint answered 400 with its real menu, which ends at `max`.
The tier was kigi's invention: the enum documented it as codex-only, the
catalog was its only producer, and the upstream model snapshot never
listed it. Removing the variant as well as the catalog entries means
`/effort ultra` is rejected at the TUI instead of parsing and failing on
the wire; a persisted `ultra` degrades to the model default through
`lenient_reasoning_effort_opt`, as designed for vocabulary changes.
kigi allowed loopback unconditionally and missed several non-public
ranges, and the SSRF check ran only on the initial URL.
Policy (ssrf.rs):
- loopback is blocked unless `[toolset.web_fetch] allow_local` (or
KIGI_WEB_FETCH_ALLOW_LOCAL) is on, AND the URL names it explicitly,
so a public name resolving to loopback stays blocked (DNS rebinding)
- add 0.0.0.0/8, 100.64/10, 192.0.0.0/24, TEST-NET-1/2/3, 198.18/15,
240/4, IPv6 site-local and documentation prefixes
- inherit the IPv4 verdict through mapped, compatible, NAT64 and 6to4
wrappers; network-specific NAT64 prefixes remain uncovered (see doc)
Plumbing (client.rs), where the exploitable half lived:
- re-check every redirect hop, not just the first
- compare hosts exactly; a `www` sibling has its own A records, so it
is a cross-host redirect rather than an auto-followed hop
- run the check before the fetch service, so a blocked URL is never
posted to an endpoint that egresses elsewhere
- exempt explicit local hosts from the https upgrade and from the
single-label filter, and re-upgrade each followed hop
Wiring: allow_local reaches WebFetchParams from both construction
paths; documented in the config guide and the README env table.
Ports grok-build's action stationarity breaker. A polling loop looks like
progress from inside the turn — every step succeeds — so nothing stops it
and it burns the whole budget re-reading one file. Nudge at 8 identical
batches, halt at 16, halt a pure no-op run at 4.
Three defects were found and fixed before this landed, none of which the
gates caught:
- The check runs AFTER execute_tool_calls, not before. Before it, the
assistant's tool_calls are recorded with no results, so pushing the
nudge triggers repair_dangling_tool_calls: synthetic "cancelled by the
user" results get spliced in, the real results land after a user
message, and dedup misses them — the orphan tool_result shape that
draws a provider 400. It is Ok-gated too, since one `?` inside can
leave a call resultless.
- `TurnOutcome::StationarityHalted` is its own variant grouped with
Completed, not a reuse of Cancelled. As a cancellation it would report
StopReason::Cancelled, fire the abort lifecycle, kill the turn's
subagents, grow the goal harness back-off streak until three halts
paused a running goal, and let completion-requirement recovery re-run
the very turn halted for looping.
- The no-op gate is size-based, not name-based. kigi registers the shell
tool as `run_terminal_cmd` and renames it `run_terminal_command`, so
matching on "bash" silently disabled the tighter ceiling while every
test still passed.
Every member was launched with the swarm's description, so the pager drew
N identical subagent blocks: the user could not tell which member was
running, and a failure could not be attributed to an item.
This is also why K6 does not port upstream's progress grid. A member IS
an ordinary subagent, so the pager already renders one block per member
with live status, duration, tool-call and turn counts — richer than a
grid cell. The gap was the label, not the widget; upstream's grid (and
the estimator behind it, which infers progress from tool-call rate)
answers a different TUI's shortcomings. The parent tool call already
reads "<description> (N members)" and `is_task_variant` already knows
`AgentSwarm`, so the blocking-wait spinner registers.
The `agent_swarm` tool works with the mode off; what the mode adds is the
doctrine — decompose finely, give every member a disjoint scope, do not
do the work yourself. `/swarm`, `/swarm on|off`, `/swarm <task>`, gated
on the tool actually being in the toolset.
Two triggers, not upstream's three. Upstream's `tool` trigger exists so
invoking the tool makes the doctrine appear; kigi's tool carries its own
description and works with the mode off, so that trigger would add a
state only the tool can reach. `Manual` persists until switched off,
`Task` expires at turn end — both are real user intents.
Load-bearing decisions, each from a defect found in review:
- Expiry is a DROP guard, not a post-loop call. A user interrupt aborts
the turn future rather than cancelling it, so post-loop code never
runs; three `?` paths skip it too. Either way a `/swarm <task>` mode
would leak into the next, unrelated prompt.
- `enter` is a total no-op while armed, so the `/swarm <task>` shorthand
cannot downgrade a standing `/swarm on` into a per-turn mode that then
disarms itself. This matches upstream; the first cut diverged, and the
test asserting the divergence was inverted with the fix.
- An explicit `/swarm off` retracts UNCONDITIONALLY. The doctrine rides
the conversation and so survives a resume, a fork and a compaction that
an in-memory flag does not; trusting the flag left a session reading
"off" with the instruction still steering and no way to clear it. That
also retires the `reminder_live` field, whose doc comment claimed to
know something the code cannot.
- The mode is deliberately NOT persisted, unlike /goal and /graph: those
strand real work when lost, this is a prompt hint whose recovery is
retyping one command. Recorded in AGENTS.md so the omission reads as a
decision, not a gap.
- The pre-session gate is `subagents_enabled`, not a hardcoded `true`:
the builder strips `agent_swarm` wherever it strips `task`, and
advertising it then offers a menu entry that resolves to literal text.
- The exit reminder is as emphatic as the doctrine it revokes; a single
weak clause is the likelier of the two to be summarised away.
The doctrine rides the existing `push_system_reminder` channel rather
than a second injection path of its own.
Ports kimi-code's AgentSwarm: a `prompt_template` containing `{{item}}`
expanded over an `items` list into up to 128 subagents, run to completion
and returned as one aggregate. Kigi already exceeds upstream on planning,
verification, isolation and merge via /graph; what it lacked was cheap
immediate fan-out. Entirely client-side — no new backend surface.
The engine is three pure pieces plus a runner: `plan` validates and
expands (every fault reported before a single member starts — a
half-launched swarm is expensive to unwind), `schedule` is the launch
ramp as testable arithmetic, `run` drives it against the existing
`SubagentBackend`. Reuses the single-spawn coordinator rather than
inventing a batch API.
Load-bearing decisions, each the result of a defect found in review:
- `backgrounded` is its own outcome. A member that outlives the 600s
foreground budget is detached by the coordinator and KEEPS RUNNING;
reporting it as failed invites the model to relaunch its item, putting
a second agent on the same files. It is never offered for resume.
- `InFlightGuard` cancels live members on Drop. Send-now cancels the turn
WITHOUT cancelling subagents and aborts the task; the dropped receivers
read as "parent gone" and each child re-attaches itself. There is no
cooperative path to use instead — `Cancellation` is constructed nowhere
in the tree — so Drop is the only seam that fires.
- Retries and wall clock are both bounded. The swarm blocks the caller's
turn, so every wait needs a ceiling it cannot argue past; stragglers at
the deadline are reported as still-running, with their ids.
- `ToolKind::AgentSwarm` is its own variant: `TemplateRenderer`'s
`by_kind` map holds one tool name per kind, so sharing `Task` would
silently redirect `${{ tools.by_kind.task }}` in other tools' prompts.
- An explicitly requested model that cannot be validated is refused, as
the task tool already does — one loud error beats `items.len()` quiet
ones. Depth stays capped at 1: upstream's unlimited nesting is a
hazard, not a feature.
- `SubagentResult.rate_limited` is classified where the typed ACP error
code is still in hand; a scheduler re-deriving it from a formatted
string would stop adapting the day the wording changed.
- Aggregate output is clamped per member (head+tail, loss stated):
native tool output is truncated nowhere downstream.
42 agent_swarm tests. The fake backend awaits, so the concurrency and
ordering assertions can actually fail; the cap test also proves the
fixture can exceed the cap.
`MemoryBackendParams` carried the primary session `AuthManager` and the
session api-key provider unconditionally, while `embed_base_url` is the
CURRENT MODEL's endpoint and `AuthRetryMiddleware` stamps `Authorization`
on every request it wraps. A user who enables `[memory.embedding] model`
while running a BYOK or subscription-OAuth model therefore sent the Kimi
session bearer to that third party.
`EndpointScopedCredentials` binds the credential to the one endpoint it
may reach: `for_endpoint` drops the handle unless the caller vouches for
the URL, and `approved_for` re-checks at provider-build time in release
too, because `MemoryBackendParams` is `Clone` and callers rewrite fields
on the copy.
The shell decides through `CredentialAuthority::manager_for` rather than
a second URL predicate — it answers both whether a credential may ride
and which manager governs it, so a subscription-OAuth platform gets its
own pooled manager. The session's `SharedApiKeyProvider` is not
forwarded at all: it is hard-wired to the primary manager, so at a
pooled platform's host it would resolve the wrong bearer. A platform's
own `embed_api_key` is untouched and keeps serving its own endpoint.
The background reindex built a second provider straight from
`ApiEmbeddingProvider::from_session`, outside the chokepoint and without
401 refresh; it now embeds through the session's own params.
Test strength verified by mutation: with the guard reverted, exactly
`session_credentials_are_withheld_from_a_foreign_endpoint` and
`a_cloned_param_set_cannot_redirect_scoped_credentials` fail.
On Windows, crossterm's console backend never emits Event::Paste: a
terminal paste arrives as a per-character key burst that the event loop
reassembles with timing heuristics. Three of its bounds split real
pastes into fragments:
- the 10 ms continue window is below one Windows scheduler quantum
(~15.6 ms), so a routine mid-burst gap ended collection early;
- the 5000-event safety cap truncated any paste past ~5 KB;
- the 2 ms detection window missed bursts whose second chunk lagged,
so collection never started.
Each fragment became its own synthetic paste. The paste chip then held
only the first fragment, and pasting again could never byte-equal the
chip content, so repaste-to-expand inserted a duplicate instead of
expanding — while macOS worked, because the unix parser buffers a
bracketed paste to the end marker and always delivers one whole event.
Widen the continue window past one quantum (25 ms), raise the cap above
any real paste (200k events), and skip the detection window entirely
when the batch already holds a >=8-key burst, which typing and
auto-repeat never produce. Unix paths are untouched: batches containing
a real Event::Paste never enter this reassembly.
Covered by a paused-clock test that trickles a burst tail with
one-quantum gaps and asserts it reassembles into a single paste.
wait_and_drain treated every read() <= 0 as "queue drained", so a real
error (bad fd, EINVAL) silently ended the drain and the monitor kept
polling a broken fd. Distinguish the cases: EAGAIN ends the drain, EINTR
retries, and any other error propagates. monitor_loop logs the error
before exiting so a dead memory monitor is visible in the session log.
Sweep every first-party crate source (1956 .rs files) to the project comment
guidelines: delete redundant restatements, decorative banners, change
narration, and end-of-line comments; keep and tighten the crucial ones
(invariants, bug rationale, SAFETY blocks, ported-source attribution).
No functional code changed. Every edit is proven comment-only against the
prior tree by a comment-stripping lexer (string/char/raw-string aware) plus a
separate doctest-fence check. Where removing a comment made rustfmt or clippy
want to re-lay-out adjacent code, the minimal triggering comment is restored so
code tokens stay byte-identical.
Gates green: cargo fmt --all --check (0 diffs), cargo check and cargo clippy
--workspace --all-targets (0 warnings).
Adds scripts/check_codegen_comment_guidelines.py — the enforcement gate for
these guidelines (flags banners, end-of-line comments, change narration, and
commented-out code).
Since v0.1.7 — chat-completions dialect correctness (Pi-referenced):
- fix(chat): BYOK/custom endpoints default to Passthrough (vanilla
OpenAI semantics; Kimi body mutations no longer leak to third-party
servers); house/Kimi coding endpoints keep the Kimi dialect via
base-url detection
- fix(chat): dedicated Mistral dialect — exactly-nine-alphanumeric
tool-call id normalizer (deterministic, call/result-symmetric)
- fix(chat): tool-result images relocate to a batched user message
(tool messages are text-only on the OpenAI chat wire)
- docs: AGENTS.md records the dialect-selection contract
C3 (Pi openai-completions relocation): tool messages are text-only on
the OpenAI chat wire — image parts inside role:tool content 400 on
strict validators. Images from a consecutive tool-result run now batch
into ONE synthetic user message appended after the run (a user message
may not interrupt tool responses answering the same assistant's
tool_calls); an image-only result carries a '(see attached image)'
pointer placeholder. Applied at the typed conversion layer, so every
chat dialect gets the universally-valid shape.
Verified: models+sampling-types+sampler+chat-state+shell all green,
clippy clean.
C1 (decision): custom/BYOK ChatCompletions entries defaulted to the Kimi
dialect, leaking Kimi-specific body mutations (thinking:{…} control,
replayed reasoning_content, schema rewrites) to arbitrary third-party
OpenAI-compatible servers. The default is now Passthrough (vanilla
OpenAI semantics), with ONE exception mirroring Pi's base-url quirk
sniffing: entries pointed at the house/Kimi coding endpoint keep the
Kimi dialect. Registry platforms are unaffected (all declare
explicitly).
C2 (Pi mistral-conversations normalizer): Mistral requires tool-call ids
of EXACTLY nine [a-zA-Z0-9] chars; even same-session synthesized UUIDs
violate it. New ChatCompat::Mistral = StrictOpenAi behavior + the
normalizer — strip non-alphanumerics, keep exact-9 ids, else FNV-1a →
base36 (build-stable, deterministic across requests for prefix-cache
stability) with collision retry; ONE map covers tool_calls[].id and
tool_call_id so pairing survives. The mistral registry row and the
persisted 'mistral' serde value both resolve to it (pre-rename Mistral
sessions gain the contract automatically); Cerebras/NVIDIA stay on
StrictOpenAi untouched.
Verified: models+sampling-types+sampler+chat-state+shell all green,
clippy clean.
Since v0.1.6 — the cross-provider replay audit (Pi transform-messages
policy: every wire builder emits only items valid for its target):
- fix(responses): reasoning items without a native rs_* id dropped (the
GPT/codex 400 Invalid 'input[N].id'); provenance gate drops foreign
reasoning and demotes foreign backend tool calls on model switches
- fix(codex): bare rs_* reasoning references dropped (stateless backend)
- fix(wire): shared ASCII tool-call id sanitizer, symmetric call+result
on both the Messages and Responses legs
- fix(messages): image-source whitelist (raster base64 only, http(s)
urls only) and empty-user-content guard
- fix(conversation): char-boundary-safe code preview (CJK/emoji code
panicked every request build)
- docs: AGENTS.md records the replay policy
Cross-provider audit M4/M5/M6 (Anthropic Messages builder):
- Non-base64 data: URIs rode as ImageSource::Url — url sources are
http(s) only → 400; and the two image paths parsed data URIs
differently (user path split on first comma, tool-result path on
';base64,').
- No media-type whitelist: image/svg+xml and param-carrying headers
('image/webp;name=x') reached the wire → 400.
- Empty user content shipped empty arrays/text blocks → 400.
One shared parse_base64_image_data_uri (raster whitelist: jpeg/png/gif/
webp) now serves both paths; rejected images degrade to a SHORT
'[unsupported image]' placeholder (never the multi-megabyte payload);
empty user turns get '[empty message]' (mirrors the assistant guard).
Part 6 of the cross-provider replay audit.
Verified: sampling-types 292 + downstream green, clippy clean.
Cross-provider audit R5 + R2-residual: BackendToolCall items round-trip
as their typed shapes with provider-issued ids (grok x_search
CustomToolCall, web_search/code_interpreter calls) and Reasoning items
carry model-bound encrypted payloads — replaying either to a DIFFERENT
Responses target names undeclared tools / undecryptable material → 400.
New transform_items_for_responses pre-pass: each [Reasoning|
BackendToolCall]* Assistant run carries provenance in
AssistantItem.model_id; on confirmed mismatch with the request's target
model, Reasoning siblings are dropped and BackendToolCall demoted to the
same text summary the Messages/ChatCompletions builders already emit.
Same-model and unknown-provenance turns stay byte-verbatim (KV-cache
prefix stability preserved). The legacy-upgrade round-trip test now
models the same-model continuation it always described.
Part 5 of the cross-provider replay audit.
Verified: sampling-types 290 + downstream 5794 green, clippy clean.
The new truncation test's doc landed between backend_tool_call_position_
stable's doc and its #[test] attribute (clippy: duplicated attribute).
Each test owns its own doc block again.
Cross-provider audit R4+M3: the Responses leg passed tool-call ids
verbatim (call_id on both function_call and function_call_output) while
the Messages leg sanitized — and its closure used Unicode
is_alphanumeric, letting CJK ids through to Anthropic's ASCII-only
contract, with no empty-id fallback. Both providers enforce
[A-Za-z0-9_-]+ (the codex 400's own words). One module-scope
sanitize_tool_call_id now serves both builders, ASCII-only, empty → "_",
applied identically on call+result so pairing survives.
Part 4 of the cross-provider replay audit.
Verified: 6081 tests green across the four crates, clippy clean.
Cross-provider audit R3: reasoning captured on stateful api.openai.com
sessions (no include requested) carries a server-issued rs_* id but NO
encrypted_content; replaying that bare reference to the stateless
(store:false) codex backend points at server state chatgpt.com does not
have. adapt_body_for_codex_backend step 3 drops such items (encrypted
ones pass through verbatim). Capture-side include for the API-key path
is deferred: the typed CreateResponse is shared with the xai Responses
leg and changing its bytes needs separate validation.
Part 3 of the cross-provider replay audit.
Verified: sampling-types+sampler+chat-state+shell all green.
The multibyte-truncation test used 80 chars — below the 100-char cap, so
the truncation assertion failed (the previous commit's suite count was
misread; the FIX itself was correct and the panic repro held). 120 chars
now exercises both the boundary safety and the truncation.
The code-interpreter preview truncated at BYTE 100 — a guaranteed panic
on any CJK/emoji boundary in interpreted code. One poisoned history item
then crashed every subsequent request build on every backend (the
summary feeds both the Messages and ChatCompletions builders). Truncate
at 100 chars via char_indices instead. Panic pinned by test.
Part 2 of the cross-provider replay audit.
Verified: sampling-types 287 tests green.
Root cause of 400 Invalid 'input[N].id': '' on chatgpt.com/backend-api/
codex/responses: the Responses input builder replayed every stored
Reasoning item verbatim, and rs::ReasoningItem.id serializes
unconditionally — so foreign items (Messages-captured Anthropic
signatures, chat-completions-synthesized reasoning, stream-delta
fallbacks, legacy upgrades — all id '') reached the wire with an empty
id the API rejects. This also self-poisoned pure codex sessions whose
reasoning arrived only as deltas.
A native item always carries a server-issued rs_* id: empty id = foreign
= unusable by any Responses provider = dropped at the builder — the
exact mirror of the Messages builder's prune_replayed_thinking. The
encrypted-only fixture that pinned the poison shape now uses a native id
(the pass-through case it always meant to cover).
Part 1 of the cross-provider replay audit (Pi transform-messages
policy: builders emit only items valid for their target).
Verified: sampling-types 286 + sampler/chat-state/shell 5791 green.
Since v0.1.5:
- fix(messages): Claude models no longer 400 with 'Invalid signature in
thinking block' — thinking is replayed only for the active tool loop
and only when genuinely signed (cross-backend/cross-model histories
are stripped)
- fix(codex): ChatGPT Codex no longer 400s with 'System messages are
not allowed' — system prompts ride the instructions field and
encrypted reasoning is requested for stateless replay
- docs: AGENTS.md records both wire contracts
Root cause of 400 {"detail":"System messages are not allowed"} at
chatgpt.com/backend-api/codex/responses (both platforms): the codex
adaptation covered only IDENTITY HEADERS (originator/OpenAI-Beta/UA/
chatgpt-account-id) — the BODY still carried the system prompt as
role:system input items, which the codex backend rejects outright. Its
system channel is the top-level field, and stateless
(store:false) reasoning replay requires
include:["reasoning.encrypted_content"] — both per the same reference
the headers were ported from (official Codex CLI + Pi's
api/openai-codex-responses.ts).
New adapt_body_for_codex_backend (kigi-sampling-types): hoists every
system input item into (order preserved, appended to any
existing instructions; string and parts content shapes) and requests
encrypted reasoning. Idempotent. Applied at both Responses send paths,
openai_codex-GATED — the API-key path stays byte-identical
(pinned by a control wire test).
Tests: adapter unit tests (hoist+include, no-system no-op, idempotence)
plus two mock-server wire tests (codex body has no system role,
instructions + include present; plain Responses body unchanged).
Verified: sampling-types + sampler + chat-state + shell 6076 tests
green, clippy clean.
Root cause of 'Invalid signature in thinking block' (400 at
messages.1.content.0, Claude models): build_messages_request replayed
EVERY stored Reasoning item as a thinking block with no origin check —
history synthesized by other backends (encrypted_content: None → the
mandatory signature field serialized as ""), Responses-API tco_* blobs
(signature bytes, no text), and blocks signed by a DIFFERENT model after
a mid-session /model switch. Anthropic validates every replayed
signature (model-bound), so such histories 400 deterministically. The
platform split was circumstantial: Windows sessions started on the
default model and switched to Claude; macOS sessions were Claude-native
from turn 1. Adversarially verified — no platform-divergent byte path
exists in capture, storage, or replay.
New prune_replayed_thinking pass (Pi/Claude Code replay policy): keep
exactly the final assistant message's thinking when its tool loop is
still open (request ends on the tool results — an open loop can never
span a model switch) and the block is genuinely signed; strip every
other thinking block (the API ignores valid prior-turn thinking and
rejects invalid). Assistant messages emptied by the strip (thinking-only
aborted turns) are removed — empty content arrays are rejected too.
Tests: three unit tests pin strip-outside-loop (unsigned, tco_*, stale
signed), keep-in-active-loop (verbatim text+signature at content.0), and
emptied-message removal; the legacy-upgrade integration test now proves
both wire fidelity in the active loop AND stripping once the loop
closes.
Verified: sampling-types + sampler + chat-state + shell 6072 tests
green, clippy clean.
The README always shipped graph engineering enabled; the enablement was
delegated to installer env plumbing that diverged per platform:
install.sh exported KIGI_GRAPH=1 into shell rc (worked), install.ps1
wrote the User registry variable — which running Windows terminals (and
new tabs of an open Windows Terminal) never pick up, so /graph was
'missing on Windows' despite a successful install.
With resolve_graph() defaulting true in the binary (e53a66d), the env
writes are redundant complexity: drop them from both installers, keep
KIGI_GRAPH=0 as the documented opt-out (env still beats the default),
and correct the flag comment to tell this story instead of a
'gray release' one. Both scripts syntax-checked (sh -n / pwsh parser).
Installers are served from main (raw.githubusercontent), so this takes
effect for all new installs immediately — no retag needed; the running
v0.1.5 build already carries the binary-default fix.
resolve_graph() read only the KIGI_GRAPH env var with default(false)
(plan.md G0 gate), so /graph existed solely on machines whose environment
exported the dev flag — which presented as '/graph is missing on
Windows'. There was never any platform-conditional code: the Mac worked
because the dev env var was set there.
Default is now true (matching /goal's shipped state); KIGI_GRAPH=0
remains the off-switch, and availability still requires the goal harness
(BuiltinGate::Graph). AGENTS.md updated; new test pins fresh-install-on
plus env-zero-off.
Verified: kigi-shell 5262 + kigi-tui 6874 tests green, clippy clean.
A single HTTP 502 from the GitHub release CDN killed the entire v0.1.5
tag build (one job of five, ~50 minutes wasted). The build-script
download now retries twice with backoff on 5xx/429/network errors;
genuine failures (404, offline) still fail fast with the
KIGI_SHELL_BUNDLE_RG_PATH hint, and every retry prints a cargo:warning
so flakiness stays observable.
GitHub Actions cache isolation lets a run restore caches only from its own
ref or the default branch. Releases run on tag refs and nothing ever ran
on main, so release.yml's rust-cache never restored anything — every
release compiled all 62 crates cold on five targets (~50 min, gated by
Windows at ~49.5 min).
warm-cache.yml builds the same release-dist profile with the same
rust-cache key on main (dep-affecting pushes — including each release's
version-bump commit — plus a weekly refresh against 7-day eviction and
manual dispatch). Tag builds then restore a warm default-branch cache;
the remaining cost is workspace-crate compilation + the codegen-units=1
thin-LTO link, which is the deliberate release-hardening tradeoff.
Since v0.1.4:
- fix(fs): Windows-safe atomic replace everywhere (util::fs::replace_file)
— model+effort switches now persist on Windows; models cache and session
state writes no longer fail silently under AV/indexer file locks
- fix(tui): a default-model persist failure keeps the live session's model
instead of reverting the pick
- docs: AGENTS.md records the replace_file contract
Persist-failure ≠ switch-failure. The default_model rollback arm re-ran
set_default_model_inner(prev) AND issued a reverse SwitchModel whenever
the config.toml write failed — the only mechanism in the codebase that
deliberately re-shows the ORIGINAL model after a successful pick. On
Windows, where AV/indexer file locks routinely failed that write (until
27d009c), every /model selection appeared to not take.
The session switch succeeds independently and reports its own failures
via handle_switch_model_complete; a disk-persist failure now keeps the
live model, logs, and surfaces only the 'Could not save' toast — the
same policy PersistPreferredModel already ships ('still active for this
session').
Verified: kigi-tui 6874 tests green, clippy clean.
Root cause of 'model+effort switch works on Mac, not on Windows': the
switch APPLIES in-session (the dispatch/apply chain is platform-identical,
verified adversarially) but its persistence never sticks on Windows.
Every tmp+rename atomic write except auth/storage.rs committed with a
bare fs::rename, and Windows MoveFileExW(REPLACE_EXISTING) fails with a
sharing violation whenever AV/search-indexer/cloud-sync transiently holds
the destination open. Consequences: [models].default never persisted
(next launch = original model), the session summary's current model never
persisted (resume = original model), and the models cache went silently
stale (all its write errors were swallowed).
- New kigi_shell_base::util::fs::replace_file — THE commit step for
tmp+rename: plain rename on Unix; on Windows delete-first + two short
backoffs (the pattern auth/storage.rs shipped first), tmp cleaned on
failure, error always returned. Windows branch type-checked against
x86_64-pc-windows-msvc.
- Adopted at every replace site: config.toml (save_config /
atomic_write_string / mcp saves), models cache (plus unique tmp
suffixes and tracing::warn on failure — writes were fully silent),
session storage (summary/current-model, jsonl, plan/signals/
announcement/goal/graph state), auth.json, active-sessions registry,
prompt history, claude/kimi import, campaigns state, goal artifacts.
Directory-move renames (worktree pool, corrupt-file backups) keep
plain rename — their destinations don't pre-exist.
Verified: kigi-shell + kigi-shell-base 5318 tests green, clippy clean,
msvc-target check of the new cfg(windows) code clean.
Since v0.1.3:
- refactor(auth): CredentialAuthority — one chokepoint for inference
credential routing; SessionCredential is unforgeable outside it
- fix(tui): login picker scrolls in a capped viewport; moon never clips
- fix(tui): each OAuth picker row routes to its own provider (LoginWith),
no more every-row-opens-kimi.com
- feat(login): /login opens the provider picker with green connected
badges; mid-session picker cancels back to the session
- fix(models): stored subscription-OAuth sessions are catalog fetch
sources — a claude-only login gets claude models and a claude default,
not the bundled Kimi table with an 'unknown' model
- feat(tui): /model + Enter opens the model picker (required-args
commands block-and-reopen); rows name their connected provider
The shell stamps meta.provider (platform display name via
parse_managed_model_key) on every managed {platform}/{model} catalog
entry in to_acp_model_info; user-defined [model.*] entries stay
provider-less. The /model dropdown surfaces it in the description column
(a model's own description still wins), so the picker reads as 'choose a
model from your connected providers'.
Verified: kigi-shell 5261 + kigi-tui 6873 tests green, clippy clean.
A required-args slash command submitted bare (/model, or /m) ran straight
into its usage error. The completeness contract already existed —
is_command_complete's documented 'Blocks' row — but had zero non-test
callers; both dispatch paths ran the command unconditionally.
Both the session prompt path and the dashboard dispatch box now consult
is_command_complete before running: an incomplete required-args command
re-opens the input as '/{command} ' with the cursor in the args phase, so
the existing suggestion dropdown lists the choices (for /model: the
catalog — by construction the connected providers' models, after
1bb10ef). Generalizes to every future required-args command; no new UI.
Verified: kigi-tui 6872 tests green, clippy clean.
Root cause of 'connected Claude, still shows Kimi / unknown model': every
fetch-plan decision consulted only the primary (Kimi) session and API keys
— never the stored subscription-OAuth sessions:
- on_auth_changed's wipe guard: a claude-pro-max-only login satisfied
'no session, no keys' → catalog wiped, early return BEFORE the fetch
and BEFORE notify_models_updated. The TUI kept an empty picker and the
prompt bar rendered 'unknown'. Guard decision extracted into the pure
should_wipe_catalog_on_auth_change (matrix-tested); a stored OAuth
session now vetoes the wipe, so the fetch runs, the first real catalog
reselects the default model (first entry = the connected provider's),
and kigi/models/update reaches the client.
- Startup prefetch: the arming gate ignored stored OAuth sessions and the
prefetch thread passed an empty token map — a claude-only user booted
onto the bundled Kimi table until a later refresh. The gate now takes
has_stored_oauth and the thread resolves each stored session's bearer
(refresh-on-expiry) via a current-thread runtime.
- Cache origins: from_config's startup cache load and cache_origin()
computed the fetch-plan origin with an empty token map, so a
claude-inclusive cached catalog never matched at startup. Both now use
presence-only stubs (stored_oauth_token_stubs — names only, no bearers)
proven equal to the real-token origin by test.
New probes in models_fetch: stored_oauth_platforms / stored_oauth_token_
stubs (sync auth.json scope scan; no AuthManager, no secrets).
Verified: kigi-shell 5260 tests green, clippy clean.
/login previously fired the resolved method's flow immediately — there was
no way to see providers or their status. It now lands on the provider
picker; already-connected providers show a green 'connected' badge in the
key column.
Shell: initialize() probes stored credentials once (primary session flag,
auth.json oauth/<provider> scopes, resolved platform keys env>auth.json>
config) and stamps _meta.connected on each advertised method
(connected_method_ids + stamp_connected_meta, pure and unit-tested).
Display state only — never an authorization input.
TUI:
- PendingMenuItem::connected() reads the badge from method meta; the
picker renders it green (accent_success), replacing the shortcut hint.
- New Action::OpenLoginPicker: /login shows the picker; mid-session it
stashes the view like dispatch_login, and starts no flow by itself.
- Mid-session picker: last row reads 'Cancel' and dispatches CancelLogin
(clicking it must not exit the app); Esc also returns to the session.
- After a successful login, the just-authenticated method is stamped
connected in the TUI's advertised-methods copy (auth_in_flight_method →
AuthComplete), so a later /login shows the badge without re-initialize.
Verified: kigi-shell 5256 + kigi-tui 6870 tests green, clippy clean.
Root cause: PendingMenuItem::Login carried only a label, so every provider
row collapsed to the id-less Action::Login, and dispatch_login resolved
the FIRST advertised interactive method — the Kimi device flow. Selecting
Grok/Claude/Copilot/Codex all opened kimi.com. (The shell side was already
correct: authenticate() dispatches each method id to its own OAuth flow.)
- PendingMenuItem::Login now carries the advertised method id (None only
on the no-interactive-method fallback row).
- New Action::LoginWith(AuthMethodId); the picker dispatches it with the
row's own id. Action::Login keeps its meaning (resolved/default method)
for /login, auto-login, and re-auth.
- dispatch_login_with resolves the id against the advertised methods and
FAILS CLOSED on an unknown id — no silent first-method fallback — then
adopts the method's label and start mode.
Tests: picker rows pinned to their method ids; LoginWith(claude-pro-max)
must authenticate with claude-pro-max even when kimi-code was previously
resolved; unknown ids surface an error and start nothing.
Verified: kigi-tui 6867 tests green, clippy clean.
Root cause (issues: unscrollable list + half-blocked moon): the stacked
welcome layout gave the menu one Length row per item. With ~30 advertised
providers the vertical layout overflowed and the constraint solver
squeezed every Length — clipping the logo — while rows past the fold were
silently unreachable.
- WelcomeLayout::compute_inner caps the menu to the rows genuinely left
over after logo/prompt/version, so the chrome never gets squeezed.
- render_menu is now a minimal-scroll viewport: it scrolls only when the
selection exits it (stable under hover — no re-centering feedback
loop), draws the picker-style scrollbar, and returns index-aligned hit
rects (zero Rect for off-screen rows) plus the offset, which AppView
feeds back next frame.
- Mouse wheel moves the login-picker selection (clamped, not wrapping).
Test: pending_menu_scrolls_and_never_clips_the_moon renders the full
method list at 40 rows and asserts the whole moon (10 braille rows), the
version badge, and that selecting Quit scrolls it into view.
Verified: kigi-tui 6864 tests green, clippy clean.
One authority answers 'which credential may ride this request':
credential_class / manager_for / credential_for / bearer_resolver_for,
keyed by (platform, base_url). SessionCredential is an opaque type with
no production constructor, so a new call site cannot re-introduce the
session-bearer leak. Platform-scoped tests extended across all bearer
channels (session, aux, summary, subagent override).
Verified: cargo check --workspace --all-targets clean; kigi-shell and
kigi-tui suites green (6611+ tests).
29th platform `openai-codex` (uses_oauth, Responses wire). PKCE-localhost login at
auth.openai.com (client app_EMoamEEZ73f0CkXaXp7hrann, redirect localhost:1455/auth/
callback, form token exchange, fresh-random state) reusing the claude-pro-max flow;
OAuthFlow::PkceLocalhost gained a redirect_path and OAuthConfig an authorize_extra
(empty elsewhere, so claude/xai/copilot authorize URLs stay byte-identical).
Codex-specific: the access token is a JWT carrying chatgpt_account_id, which becomes
the `chatgpt-account-id` inference header. It is derived STATELESSLY from whichever
bearer rides each request (so a rotated token needs no persisted field), and BOTH
login and refresh fail fast when the claim is absent — gated on the explicit
OAuthConfig.requires_chatgpt_account_id fact, never inferred from the token-body
encoding (a plain form endpoint is the OAuth norm and must not inherit this).
Inference rides the existing Responses wire at chatgpt.com/backend-api/codex →
/responses, with codex headers (chatgpt-account-id, originator, OpenAI-Beta
responses=experimental, codex UA) gated on SamplerConfig.openai_codex so API-key
`openai` stays byte-identical; store:false was already the global Responses default.
Catalog is HARDCODED (no live endpoint exists for this backend; read from the
official Codex CLI's model cache): gpt-5.6-sol/terra/luna + gpt-5.5, ctx 272000,
each with its real reasoning levels (low..ultra — ReasoningEffort gained Ultra).
Excluded: gpt-5.3-codex-spark (supported_in_api=false), gpt-5.4/-mini and
codex-auto-review (hidden) — they would list but fail at inference. The fetch
short-circuits before any HTTP; Kigi never shells out to the codex CLI or reads
~/.codex.
Security review fixes: redact any `account-id` header from request logs (it was
reaching debug logs), strict 3-segment JWT check (fail closed), refresh no longer
fails open on a missing claim. Inherits leak-safe pooled routing (scope
oauth/openai-codex) — never the Kimi token. Full gate green (234 suites, 0 warnings).
28th platform `github-copilot` (uses_oauth, ChatCompletions wire). Two-stage auth:
RFC-8628 GitHub device flow (client Iv1.b507a08c87ecfe98, scope read:user, errors
in a 200 body) mints the DURABLE github token; a GET api.github.com/copilot_internal/
v2/token exchange re-mints the SHORT-LIVED copilot token. Persisted as key=copilot
token, refresh_token=github token, expires_at=copilot expiry; the "refresh" is a
copilot-token re-mint (not a refresh_token grant), dispatched via
OAuthTokenBody::GithubCopilotExchange in the generic refresher.
VS Code editor-identity headers on /models + /chat/completions, gated on
SamplerConfig.github_copilot / PlatformId::sends_copilot_editor_headers() so every
other ChatCompletions provider stays byte-identical. Live /models filtered
(parse_github_copilot_listing) to the openai-completions-served models: keep iff
model_picker_enabled && policy.state!="disabled" && tool_calls!=false AND not a
claude-4.x/5.x (messages) or gpt-5/oswe/mai- (responses-only) id — those need
per-model wire routing (documented debt), excluded rather than mis-routed.
Inherits the leak-safe pooled routing (scope oauth/github-copilot); its bearer/
refresh/api_key never touch the Kimi token (regression test added). Fail-fast on
an out-of-range copilot expires_at (would otherwise silently 401 mid-session).
Adversarial security review: GO, no CRITICAL/HIGH. Known limitation: Pi's
per-model policy-enablement POST is not ported (documented in AGENTS.md).
27th registry variant, 2nd subscription-OAuth provider. Log in with a Claude
Pro/Max subscription via PKCE authorization-code + S256 (loopback callback on
127.0.0.1:53692, with a manual code-paste fallback), then use it against
api.anthropic.com — reusing the existing Anthropic Messages wire + Anthropic
listing + the multi-provider OAuth foundation (dbce6bf). Sourced from Pi
(earendil-works/pi auth/oauth/anthropic.ts): client 9d1c250a..., authorize
claude.ai/oauth/authorize, token platform.claude.com/v1/oauth/token, scope
'…user:inference user:sessions:claude_code…'.
New machinery (foundation handles token routing — claude-pro-max is a
uses_oauth platform so its bearer/refresh/api_key already route to its own
pooled manager, never Kimi):
- OAuthConfig gains flow{DeviceCode|PkceLocalhost} + token_host + token_body
{Form|JSON}; xai/kimi rows unchanged (DeviceCode/Form).
- auth/oauth_pkce.rs: PKCE S256 wire — loopback listener with STRICT state
validation (CSRF, fail-closed), manual-paste fallback, JSON code→token
exchange + rotating-refresh. Never logs code/verifier/tokens.
- Messages OAuth adaptation gated on SamplerConfig.anthropic_oauth (true only
for a claude-pro-max managed key): Authorization: Bearer + anthropic-beta
oauth + user-agent claude-cli + x-app cli, and the required 'You are Claude
Code' system prefix. API-key anthropic/minimax Messages requests are
BYTE-IDENTICAL (regression-guarded).
- Live /models under the OAuth Bearer + oauth-beta headers (Anthropic listing,
enriched from models.dev anthropic); persistent 401 → 0 models + WARN, NO
hardcoded fallback list (honest failure).
Adversarial review: no blocking findings (secret handling, CSRF/state, the
anthropic_oauth gate, token routing, non-regression all CONFIRMED). Full gate
green. Registry at 27; picker updated. Residual (unverifiable without a real
Claude Pro/Max account): whether GET /v1/models accepts the OAuth bearer, and
the real endpoint's acceptance of the OAuth Messages request.
First subscription-OAuth provider beyond Kimi Code (26th registry variant).
Log in with a Grok/SuperGrok/X subscription via RFC-8628 device-code OAuth
(auth.x.ai), then use it against api.x.ai/v1 — reusing the existing xai wire
(ChatCompletions + OpenAI listing + Passthrough + restrict + models_dev_id
xai). Sourced from Pi (earendil-works/pi auth/oauth/xai.ts): client
b1a00492..., scope 'openid profile email offline_access grok-cli:access
api:access', standard Bearer (no x-xai-token-auth).
Foundation (generalizes Kigi's Kimi-singleton OAuth to per-provider, root
cause, not a patch):
- Registry: OAuthConfig on PlatformSpec (client_id/host/device+token
paths/scope/scope_key); XAI_OAUTH_CONFIG + XAI_GROK_SPEC (uses_oauth, method
id 'xai-grok', an interactive login after kimi-code).
- Generic device-code wire (auth/oauth_device.rs) + GenericDeviceRefresher,
sharing the RFC-8628 core with Kimi; Kimi's bespoke flow is byte-identical
(X-Msh headers, KIMI_CODE_OAUTH_SCOPE, keyring gating unchanged).
- Per-provider AuthManager via a process-global pool (auth/oauth_registry.rs):
build-on-demand with start_proactive_refresh, keyed by scope. The session
resolves the AuthManager for the ACTIVE model's platform for bearer/refresh/
401-recovery/api_key — an oauth-platform model always uses its OWN token,
never the primary.
- Live /models under OAuth; base routes oauth().is_some() -> platform.base_url()
(kimi-code stays on proxy_url).
Security: adversarial review + a systematic token-leak audit found and closed
FIVE channels where the primary Kimi token could reach api.x.ai (bearer
resolver, api_key stamping, aux summary/classifier/image-describe models, and
subagent model-override). Each fix routes through the platform-aware resolver
(the oauth model's pooled token or None, NEVER the primary) and is revert-to-red
verified. No access/refresh token is ever logged.
Registry at 26; picker updated (xai-grok interactive login row); TUI
context-window already auto-updates per model. Full gate green (234 suites,
fmt, clippy -D warnings, deny). GPT/Claude/Grok officially permit third-party
subscription use.
Kigi started as an unofficial Kimi Code CLI community build; by request it
now works with 25 providers. Replace the stale three-platform table with the
full supported list (Kimi Code OAuth + 24 API-key providers, each with its
platform id and env var). Unsupported/not-yet-built providers are omitted.
Providers 21-22 (24th & 25th registry variants), sourced from Pi
(earendil-works/pi). Pi drives MiniMax through its Anthropic-COMPATIBLE
surface (baseUrl .../anthropic), so Kigi reuses the existing Anthropic
Messages machinery (wire_api=Messages, listing=Anthropic, key_header=XApiKey
x-api-key+anthropic-version) rather than the OpenAI path. Global:
api.minimax.io/anthropic, MINIMAX_API_KEY, models.dev minimax. China:
api.minimaxi.com/anthropic, MINIMAX_CN_API_KEY, models.dev minimax-cn.
The base carries the /v1 suffix (.../anthropic/v1) since Kigi appends bare
paths → listing .../anthropic/v1/models?limit=1000, inference
.../anthropic/v1/messages (matches the live-probed x-api-key-gated endpoint).
restrict_to_enriched=FALSE: the 7 MiniMax-M* models are clean (no pollution)
and restrict would drop launch-day models not yet in models.dev.
Also HARDENS parse_anthropic_listing to tolerate a bare array in addition to
the {data:[...]} envelope (mirrors parse_openai_listing's sniff that Together
taught us) — so MiniMax's Anthropic-compatible /models can't silently empty
the catalog if it serves a bare array. A bare object without data still errors.
Review found no defects (5 areas CONFIRMED incl. the /v1 non-doubling, the
additive parser change, restrict=false rationale). Residual (logged): the live
200 body / anthropic-version acceptance is unverifiable without a key; the
bare-array tolerance + restrict=false hedge most shapes.
Tests: e2e mocks the x-api-key-gated Anthropic listing (proves the auth header
+ enrichment + keying under minimax/); both validation tests reject 401 with
the per-variant console host; new parser test covers envelope + bare array +
bare-object-errors. Registry at 25; picker 26 rows.
Providers 19-20 (22nd & 23rd registry variants), sourced from Pi
(earendil-works/pi). Pi's xiaomi.ts / xiaomi-token-plan-cn.ts use plain
openAICompletionsApi (no thinking dialect) → Kigi Passthrough. Global: MiMo
at api.xiaomimimo.com/v1, XIAOMI_API_KEY, models.dev xiaomi. China token plan:
token-plan-cn.xiaomimimo.com/v1, XIAOMI_TOKEN_PLAN_CN_API_KEY, models.dev
xiaomi-token-plan-cn. Both Bearer + OpenAI listing + ChatCompletions +
restrict_to_enriched; /models auth-gated → validator.
id-match proven vs Pi's static ids (mimo-v2.5-pro, mimo-v2-flash, ...). The
CN token plan lists 4 mimo TTS models (tool_call=false); restrict's tool_call
cut drops them, keeping only the 3 chat models — the e2e proves this
non-vacuously with the real mimo-v2-tts. Review found no defects.
Registry at 23; picker 24 rows (welcome picker verified rendering all rows at
a taller viewport).
Providers 17-18 (20th & 21st registry variants), sourced authoritatively
from Pi (earendil-works/pi), the open-source agent whose provider list is
being mirrored. Pi's zai.ts / zai-coding-cn.ts use plain openAICompletionsApi
(NO special thinking dialect — overturns the matrix's 'thinking:{type} → new
dialect' concern), so Kigi maps them to Passthrough. Global: api.z.ai/api/
coding/paas/v4, ZAI_API_KEY, models.dev zai-coding-plan. China (Zhipu
BigModel): open.bigmodel.cn/api/coding/paas/v4, ZAI_CODING_CN_API_KEY,
models.dev zhipuai-coding-plan. Both Bearer + OpenAI listing + ChatCompletions
+ restrict_to_enriched; /models is auth-gated → validator.
id-match is PROVEN (not just assumed like Qwen): Pi's static model ids
[glm-4.5-air, glm-4.7, glm-5-turbo, glm-5.1, glm-5.2, glm-5v-turbo] are
byte-identical to the models.dev zai-coding-plan keys, all tool_call=true, so
restrict keeps every model with no silent-empty risk. Review found no defects.
GLM thinking is not lost (reasoning_content is parsed regardless of dialect).
Tests: e2e proves enrichment-supplied context + non-vacuous restrict (a
non-enriched wire model is dropped) + Passthrough; both variants' validation
tests hit /models and assert the per-variant console host (z.ai vs
open.bigmodel.cn). Registry at 21; picker 22 rows.
Also fixes the welcome login-picker test for the now-taller menu (renders at
a taller viewport to verify content coverage) and logs the real menu-overflow
UX debt: the picker clips rows past the fold with no scroll (q/l shortcuts
still work; only shown when unauthenticated) — deferred to its own cycle.
Provider 16 (19th registry variant). Same endpoint + models + Kimi dialect
as the existing OAuth kimi-code platform (api.kimi.com/coding/v1 via the
KIGI_CODE_BASE_URL override), but authenticated with a static KIMI_API_KEY
instead of the device flow — for users who have a Kimi For Coding key rather
than an OAuth subscription. Bearer, OpenAI listing, ChatCompletions,
ChatCompat::Kimi, wire_serves_metadata=true (Kimi /models self-serves
context/thinking), restrict_to_enriched=false (clean 3-model catalog).
/coding/v1/models is auth-gated (401) so it doubles as the validator.
No collision: KIMI_API_KEY was previously unused (grep-verified), and the
house BYOK reads only KIGI_API_KEY/XAI_API_KEY/legacy. kimi-code (OAuth) and
kimi-coding (static key) are independently gated (OAuth-token vs key) and
their models get distinct managed keys (kimi-code/k3 vs kimi-coding/k3) — a
user with both simply sees each Kimi model twice; no dedup collision, no crash.
Review (6 areas): no blocking defects; confirmed the spec correctly mirrors
KIMI_CODE_SPEC (differing only in uses_oauth/api_key_envs/console_host/labels)
and the coexistence is benign. Strengthened the e2e's dialect assertion (Kimi
is the default ChatCompat, so it did not discriminate a parse failure) by
also asserting parse_managed_model_key attributes the key to KimiCoding.
Tests: e2e proves wire-served context (1_048_576 from the wire) with the
models.dev fetch SKIPPED (all-wire-metadata provider, .expect(0)), bare-id
round-trip under kimi-coding/, Kimi dialect; validation test rejects a 401
from /models. Registry at 19; picker 20 rows; snapshot already bundles
kimi-for-coding.
Providers 14-15 (17th & 18th registry variants), Alibaba DashScope
compatible-mode. Global: token-plan.ap-southeast-1.maas.aliyuncs.com,
QWEN_TOKEN_PLAN_API_KEY, models.dev alibaba-token-plan. China:
token-plan.cn-beijing.maas.aliyuncs.com, QWEN_TOKEN_PLAN_CN_API_KEY,
alibaba-token-plan-cn. Both Bearer + OpenAI listing + ChatCompletions +
Passthrough (stream_options.include_usage is documented-supported).
/models is auth-gated (401 without a key) so it doubles as the validator;
metadata from models.dev enrichment. The token plan is a multi-vendor
catalog (deepseek/kimi/minimax/glm/qwen); restrict_to_enriched keeps the 15
tool-calling chat models and drops the 4 qwen-image/wan image generators.
Review downgraded two flagged concerns: the enable_thinking non-streaming
400 cannot occur (Kigi never issues non-streaming ChatCompletions in
production — all inference streams), and Qwen thinking is NOT invisible
(reasoning_content is parsed regardless of the Passthrough dialect). No
defects. Residual (logged): restrict does an exact id-match of live /models
ids vs the models.dev keys; a mismatch fails safe (0 models) — verify with a
real key. Snapshot already bundles both providers (regenerated in ebf1105).
Tests: global e2e proves enrichment-supplied context (wire carries none),
non-vacuous tool_call restriction (qwen-image dropped), bare-id round-trip,
Passthrough; both variants' validation tests hit /models (401 reject) and
assert the correct per-variant console host. Registry at 18; picker 19 rows.
13th provider (16th registry variant). Also reconciles a naming collision
the matrix flagged: this fork is house-branded "xai" (cf. xai.dev metadata,
KIGI_CODE_XAI_API_KEY legacy env), so XAI_API_KEY + method xai.api_key were
the GENERIC house BYOK, not x.ai/Grok. The provider table wants xai/XAI_API_KEY
for Grok.
Resolution (user-approved): XAI_API_KEY now keys the x.ai/Grok provider; the
house BYOK primary env moves to KIGI_API_KEY, keeping XAI_API_KEY and
KIGI_CODE_XAI_API_KEY as back-compat fallbacks (read_xai_api_key_env checks
KIGI_API_KEY first). The xai.api_key method id is unchanged (persisted-session
compat); the platform method id is the bare "xai", distinct from it.
xAI spec: api.x.ai/v1, Bearer, OpenAI listing + ChatCompletions, Passthrough
(docs confirm stream_options.include_usage accepted). /v1/models is minimal
(ids only) and requires auth, so it doubles as the key validator (401 on bad
key, no override) and metadata comes from models.dev enrichment. Live ids
match the models.dev "xai" keys byte-for-byte, so restrict_to_enriched keeps
the 5 tool-calling chat models (grok-4.5/4.3/4.20-0309-*/build-0.1) and drops
the grok-imagine-* generators + the non-tool multi-agent model. Snapshot
regenerated to include the xai provider (was stale; gen script already listed
it in TARGETS).
Env migration is comprehensive to avoid keying the xai platform (which would
trigger a live api.x.ai fetch) or leaving house-key reads stranded: routed
the trace CLI resolver + acp_agent/auth.json bridge + paste-key ext handler
through the new primary; moved all leader/pager/e2e harness setters to
KIGI_API_KEY; made every house-key isolation test unset KIGI_API_KEY too;
updated user-facing hints to name KIGI_API_KEY.
Tests: e2e proves enrichment-supplied context (wire carries none), non-vacuous
tool_call restriction, bare-id round-trip under xai/, Passthrough; validation
tests hit /models (401 reject, 200 accept); house_env_var_takes_precedence_over_xai
pins the new precedence. Registry at 16; picker 17 rows; 4 auth arrays + xai.
Review (16 findings, all fixed): caught a missed else-branch env clear in the
paste-key handler (would leak the house key past a clear) and a non-hermetic
credential-priority test; both fixed.
12th provider. API-key via AI_GATEWAY_API_KEY, Bearer, OpenAI listing +
ChatCompletions, Passthrough dialect. Second wire-metadata provider but
takes the enrichment path instead: Vercel serves context under
context_window, which WireModel ignores (reads context_length), so
wire_serves_metadata=false + restrict_to_enriched pulls context/limits
from the models.dev "vercel" snapshot (302/306 live ids match snapshot
keys byte-for-byte, so restrict keeps essentially the whole catalog).
/models is public (200 for any key), so login validation targets
/credits (key_validation_path) which 401s on a bad bearer — avoids
false-accepting invalid keys against the public listing.
Tests: e2e proves enrichment-wins (wire context_window=999 distinct from
enrichment context=400000, asserts 400000) and non-vacuous tool_call
restriction; validation test proves /credits (not /models) is hit.
Registry at 15 (ordinal/VARIANT_COUNT/ALL), 4 auth arrays + 16-row picker.
The 14th registry row: id "nvidia", NVIDIA_API_KEY > auth.json "nvidia"
scope, https://integrate.api.nvidia.com/v1 with KIGI_NVIDIA_BASE_URL
override, Bearer, ChatCompletions, enrichment-backed metadata
(models_dev_id nvidia — the quirk matrix's earlier 'absent' claim was
wrong; models.dev has 84 nvidia models), restrict_to_enriched=true (the
NIM listing mixes chat/embedding/rerank/vision/image; keep the 45
tool-calling chat models). Slashed org/model ids
(nvidia/meta/llama-3.3-70b-instruct) round-trip via the first-slash split;
the native id rides the wire.
NIM exposes raw vLLM behavior and stream_options support varies per model
(some strict vLLM backends 4xx on it), so chat_compat=StrictOpenAi strips
stream_options — streaming works across the whole fleet, usage falls back
to estimation. Snapshot reasoning models carry no effort menus, so kigi
sends no reasoning_effort (which an unsupported strict validator would
400 on). Review: no defects. Logged note: a key lacking the org 'Public
API Endpoints' permission passes /models validation but 403s on chat
(user-fixable edge case).
The 13th registry row: id "cerebras", CEREBRAS_API_KEY > auth.json
"cerebras" scope, https://api.cerebras.ai/v1 with KIGI_CEREBRAS_BASE_URL
override, Bearer, ChatCompletions, enrichment-backed metadata
(models_dev_id cerebras).
Cerebras' catalog is all chat LLMs (no embedding/tts pollution) and its
/models is minimal (ids only), so restrict_to_enriched=FALSE: keep every
live model, enrich the known ones (context + effort menus low/medium/high),
unknown ones keep the default context. The e2e pins this enrich-without-
restrict path (new — prior enrichment providers all used restrict=true).
Review caught a likely-DOA defect: Cerebras uses strict
additionalProperties:false validation (confirmed 400-rejecting store,
maxTokens, thinking, nested reasoning_content), and stream_options is not
in its schema — so Passthrough (which keeps the stream_options.include_usage
kigi injects on every streaming request) would very likely 400 all
streaming. Generalized ChatCompat::Mistral -> ChatCompat::StrictOpenAi
(serde alias "mistral" keeps pre-rename persisted sessions loading), which
strips stream_options + private fields for any strict OpenAI-compat
validator; both Mistral and Cerebras now map to it. Future strict-validator
candidates (NVIDIA/Azure/Xiaomi/OpenCode) noted for the same check.
reasoning_effort (incl. "none") passes through; /v1/models requires auth
so key validation works; console cloud.cerebras.ai.
The 12th registry row: id "together", TOGETHER_API_KEY > auth.json
"together" scope, https://api.together.xyz/v1 with KIGI_TOGETHER_BASE_URL
override, Bearer, ChatCompletions + Passthrough, enrichment-backed
metadata (models_dev_id togetherai) with the tool-calling listing
restriction (Together's listing mixes chat/embedding/rerank/image types).
Together's GET /v1/models returns a BARE JSON ARRAY, not the OpenAI
{object:list,data:[]} envelope. New shared parser parse_openai_listing
tolerates both shapes (sniffs the top-level [ vs { for accurate
diagnostics); every OpenAI-listing provider now routes through it, with
byte-equivalent envelope behavior (verified: Groq/Google/OpenRouter
unchanged) and non-silent errors. Together's org/Model ids match the
togetherai snapshot keys exactly (verified), so restrict_to_enriched keeps
the ~27 tool-calling models without the id-shape trap.
Review: ship-ready, shared-parser change proven strictly-additive and safe,
registry/e2e/dialect all correct. Backlog logged: kigi ignores the wire
per-model field, so a live Together chat model absent from
models.dev is dropped until indexed — a future wire- filter would
unlock the fresh full catalog.
The 11th registry row and the first THIRD-PARTY wire_serves_metadata=true
provider: id "openrouter", OPENROUTER_API_KEY > auth.json "openrouter"
scope, https://openrouter.ai/api/v1 (note /api/v1) with
KIGI_OPENROUTER_BASE_URL override, Bearer, OpenAI listing +
ChatCompletions + Passthrough.
OpenRouter's public /models serves context_length for every model
(verified live: 340/340), so it needs NO enrichment: models_dev_id=None,
wire_serves_metadata=true, restrict_to_enriched=false. An OpenRouter-only
user makes zero models.dev calls; context comes straight from the listing.
Slashed ids (anthropic/claude-opus-4.8) round-trip through the managed key
via the first-slash split; the native id rides the wire. The e2e pins all
of this with the models.dev refresh disabled.
Review-confirmed defect fixed (and independently re-verified with live
curls): OpenRouter's /models is PUBLIC — GET /models returns 200 for ANY
key — so login key-validation would false-accept a bad key, deferring the
failure to the first chat 401. New spec field key_validation_path lets a
public-listing platform validate against an auth-requiring endpoint;
OpenRouter uses /key (401s for bad keys). Reusable for Vercel (also
public). Tests pin the /key validation and no regression to the default
/models path.
Gate caught a fixture regression: the kimi_import test used openrouter.ai
to represent a CUSTOM provider, which now correctly dedupes to the builtin
OpenRouter — moved the fixture to a reserved llm.example.test host that no
future platform can shadow.
The 10th registry row: id "google", GEMINI_API_KEY > auth.json "google"
scope, https://generativelanguage.googleapis.com/v1beta/openai (Gemini's
OpenAI-compatibility shim) with KIGI_GOOGLE_BASE_URL override, Bearer,
OpenAI listing + ChatCompletions + Passthrough, enrichment-backed
metadata (models_dev_id google), tool-calling listing restriction.
Review caught a ship-blocking defect: Gemini's compat /models returns
-PREFIXED ids (confirmed via Google's own cookbook), but the
models.dev snapshot keys and the chat endpoint use the BARE id. Without
normalization the enrichment lookup misses and restrict_to_enriched
silently empties the Gemini catalog (login works, zero models
selectable). A doc WebFetch had hidden this — the compat docs show bare
INPUT ids to retrieve/chat but never print the list OUTPUT.
Fix: new spec field strip_listing_id_prefix, applied in the fetch before
filter/enrich/keying. Google sets Some("models/") (defensive: a no-op if
an id is already bare, so correct regardless of the live shape); all other
rows None. The e2e now feeds the REAL prefixed listing ids and asserts
they survive as the bare managed key google/gemini-2.5-pro with the bare
id on the wire (chat rejects the prefix); a registry test pins the config.
Note: models.dev models Gemini reasoning as budget_tokens (not
effort-type), so no auto effort-menu — models still fully work; reasoning
is dynamic. Gemini compat applies default safety filters (no BLOCK_NONE).
The 9th registry row, pure Groq pattern: id "fireworks",
FIREWORKS_API_KEY > auth.json "fireworks" scope,
https://api.fireworks.ai/inference/v1 (note /inference/v1) with
KIGI_FIREWORKS_BASE_URL override, OpenAI listing + ChatCompletions +
Passthrough dialect, enrichment-backed metadata (models_dev_id
fireworks-ai) with the tool-calling listing restriction.
Fireworks native ids are deeply slashed (accounts/fireworks/models/glm-5p2);
the e2e pins the full round-trip: the managed key
(fireworks/accounts/fireworks/models/glm-5p2) parses back on the first
slash, and — the 404-risk property — the NATIVE id rides the inference
wire (entry.model and the resolved SamplerConfig.model) while the
fireworks/ prefix stays internal routing only.
Review (models.dev provider.toml + Fireworks docs): row facts confirmed,
registry integrity at 9, counts complete, e2e strong on all axes, zero
defects. One tradeoff logged as debt: restrict_to_enriched drops
fine-tuned/account-scoped deployed models (a headline Fireworks feature)
that can never be in models.dev.
The 8th registry row: id "mistral", MISTRAL_API_KEY > auth.json
"mistral" scope, https://api.mistral.ai/v1 with KIGI_MISTRAL_BASE_URL
override, enrichment-backed metadata with the tool-calling listing
restriction (embed/moderation/OCR noise).
Mistral is NOT a pure-pattern provider — an adversarial review found two
doc-confirmed blockers that no test exercises (no e2e covers a chat POST),
so the gate-green registry row alone would have shipped it DOA. A research
workflow pinned the exact wire shapes against the mistralai/client-python
SDK source (adversarially verified), then both were fixed:
1. stream_options 422: Mistral's strict Pydantic validator rejects the
stream_options.include_usage field kigi injects on every streaming
request (the SDK's request model has no such field). New
ChatCompat::Mistral dialect strips it (plus the kigi-private message
fields, like Passthrough). Streaming usage falls back to token
estimation.
2. Reasoning content arrays: Mistral reasoning models return content as
Union[str, List[ContentChunk]] on both streaming and non-streaming,
which the flat Option<String> path could not decode -> aborted turn.
A UNIVERSAL lenient deserializer (#[serde(from = "Raw..")] on
ChatResponseMessage + ChatChunkDelta) accepts string-or-array, routing
{type:text} chunks to the answer and the nested text of {type:thinking}
chunks to reasoning_content, tolerant of the OPEN chunk union (unknown
types ignored, never fatal). String content stays byte-identical for
every other provider (kimi/deepseek/groq/BYOK).
Review refuted all seven attack lines (no regression, no crash, exhaustive)
and flagged one coverage gap, now closed: a stream-consumer integration
test drives a full thinking -> transition -> answer chunk sequence and
proves it yields the same reasoning-sibling + assistant-answer result as
the reasoning_content string path.
Also folds a verified quirk matrix for all 23 remaining API providers into
providers-plan.md, tiered by real difficulty (self-enriching OpenRouter/
Vercel; bare-array Together listing; Messages-dialect MiniMax reusing the
Anthropic machinery; non-Bearer Azure/Bedrock; router wildcards; the OAuth
block).
The 7th registry row and the first pure-pattern cycle: id "groq",
GROQ_API_KEY > auth.json "groq" scope, https://api.groq.com/openai/v1
with KIGI_GROQ_BASE_URL override, OpenAI listing + ChatCompletions +
Passthrough dialect (Groq accepts the OpenAI-style reasoning_effort
scalar verbatim), enrichment-backed metadata with the tool-calling
restriction (the listing carries whisper/tts/guard noise — 8 of 15
enrichment entries are non-chat).
Review verdict: faithful pattern repeat, zero blocking findings — row
facts verified against live Groq docs, registry integrity at 7, e2e
proven strong on both axes (restriction + dialect mapping). Added the
recommended pin: managed keys split on the FIRST slash, so Groq's
provider-native slashed ids (openai/gpt-oss-120b and 10 more) round-trip
as groq/openai/gpt-oss-120b → (Groq, openai/gpt-oss-120b).
Also fixes a pre-existing test race this cycle surfaced: the enterprise-
endpoints test asserted moonshot's fixed base while non-serial, racing
serial tests that legitimately point KIGI_MOONSHOT_CN_BASE_URL at
wiremock; now serial + env-unset like its documented siblings. Two
display-only advisories logged as tracked debt (Groq's delta.reasoning
field invisible in the TUI; platform-generic rate-limit copy).
The 6th registry row: id "deepseek", DEEPSEEK_API_KEY > auth.json
"deepseek" scope, base https://api.deepseek.com (chat rides
{base}/chat/completions per official docs) with KIGI_DEEPSEEK_BASE_URL
override, enrichment-backed metadata (1M context, 384k output cap,
high/max effort menu).
Structural fix the cycle exposed: kigi's Kimi-specific body adaptation
ran UNCONDITIONALLY on every ChatCompletions request. New ChatCompat
dialect, declared per platform row and threaded through SamplerConfig,
ClientDefaults, and the session-persisted SamplingConfig (serde-default
Kimi keeps restored pre-field sessions and BYOK endpoints byte-identical;
production persist seams copy it; subagents inherit it):
- Kimi: full legacy pipeline (dispatch ≡ legacy pinned)
- DeepSeek: thinking:{type, reasoning_effort} per api-docs.deepseek.com
(server maps low/medium→high, xhigh→max itself; none disables; absent
leaves the server default)
- Passthrough: OpenAI-style reasoning_effort scalar untouched (unblocks
Groq and the rest of the OpenAI-compatible list)
Review-confirmed release blocker fixed: kigi replays Kimi's
reasoning_content (and its private model_id) on input assistant
messages — Kimi consumes these, but DeepSeek documents input
reasoning_content as prefix-mode-only (historically a 400) and other
providers don't know either field. The DeepSeek and Passthrough arms now
strip both; Kimi's own pipeline is untouched. Pinned on both message
shapes.
The 5th registry row: id "anthropic", ANTHROPIC_API_KEY > auth.json
"anthropic" scope, api.anthropic.com/v1 with KIGI_ANTHROPIC_BASE_URL
override, Messages dialect. Two new spec dimensions most future rows
reuse: ListingDialect (Anthropic's /v1/models wants x-api-key +
anthropic-version headers, ?limit=1000, and its own response shape) and
PlatformKeyHeader (Bearer vs x-api-key across listing/validation/
inference, with auth_scheme stamped onto entries).
The 2026 Anthropic listing serves real metadata: the adapter maps
max_input_tokens, per-level effort capabilities (low..max as the menu,
xhigh/max distinct), thinking/image flags — and enrichment fills only
genuine wire gaps (e2e pins wire-1M beating enrichment, and a zero
context filled to 200k).
Two review-confirmed defects fixed red-green:
- Output caps were dropped at three layers, so every sub-128K-output
model (64k Haiku, legacy models) would 400 on EVERY request against
the sampler's 128K max_tokens default. Wire max_tokens and enrichment
limit.output now flow to entry.max_completion_tokens.
- An explicit wire effort-decline was indistinguishable from wire
silence, letting enrichment inject effort menus pre-4.6 models reject
(adaptive thinking 400). The adapter now emits a decline sentinel
(support:false) that enrichment respects — proven end to end.
Also: the Messages client now sends anthropic-version (previously never
sent — real api.anthropic.com rejects such requests; pinned across all
three scheme/backend quadrants), key validation builds per-key-header
requests, missing listing data fails fast, empty-id ghosts drop with a
warning, kimi-import recognizes api.anthropic.com as built-in
automatically.
The 4th registry row: id "openai", OPENAI_API_KEY env > auth.json
"openai" scope (login picker/paste/validation all registry-generic —
zero TUI changes needed, pinned by the picker test), base
https://api.openai.com/v1 with KIGI_OPENAI_BASE_URL override, Responses
dialect via the new PlatformWireApi spec field, enrichment-backed
metadata (wire_serves_metadata=false).
OpenAI's GET /v1/models returns bare ids and is polluted with
tts/whisper/embeddings entries: the listing is restricted to
enrichment-known TOOL-CALLING models (review caught that membership
alone admitted models.dev-known embeddings models, which would 400 on
every agentic request; dropped ids are debug-logged for launch-day
diagnosability). Context windows, effort menus, display names, and
thinking capability come from the enrichment pipeline — wiremock e2e
pins the full contract: polluted live listing + models.dev →
one Responses-backed chat model with a 400k documented context window.
Responses max-effort wiring (closes the P0c-1 debt): canonical effort
rides a CreateResponseWrapper sidecar and patch_reasoning_effort writes
it onto the serialized body at both send sites (all seven levels pinned,
xhigh/max distinct, summary preserved); normalize_effort_echo drops
echoes async-openai's typed enum cannot represent at both the non-stream
and SSE parse seams; the dead typed to_responses_api converter is
deleted. Kimi/moonshot stay byte-identical (ChatCompletions untouched,
wire_api maps to the same default; kimi wire tests green).
kimi-import now recognizes ANY registry platform host as built-in
(was hardcoded moonshot), covering openai and future rows.
Provider /models listings that return bare ids (OpenAI-style) get context
windows, thinking levels, image support, and display names from models.dev:
kigi-models owns the transform (parse_api_json — ONE field interpretation
for the bundled snapshot AND runtime refreshes), enrich_wire_model fills
gaps with wire values always winning and model availability strictly
wire-truth. Spec rows gained models_dev_id + wire_serves_metadata; all
three current platforms are wire-served, so this pipeline is provably
inert for them (byte-identical catalogs, zero egress, zero ~/.kigi writes
— adversarially verified).
Shell side: enrichment_fetch with a 24h disk cache guarded by binary
version + keep-set + future-stamp sanity (a registry change or downgrade
refetches instead of serving a catalog missing new providers), refresh of
https://models.dev/api.json filtered to registry ids, KIGI_MODELS_DEV_URL
override with case/whitespace-tolerant kill switch, fallback chain
fresh-cache > refresh > stale-cache > bundled (each step logged). The
fast path returns an empty catalog without forcing the bundled parse.
From the review: blast-radius-confined parsing (one drifted provider on
models.dev warn-skips instead of failing the whole refresh), registry-
coverage and field-coverage tests guarding script/parser drift, a path-
injectable core with 8 state-machine tests (one of which caught a guard
patch that had failed to apply), _meta provenance stamp in the snapshot,
and models.dev (MIT) attribution in NOTICE. Snapshot: 29 providers, 1124
models, 246KB, regenerated by scripts/gen_enrichment_snapshot.py (pure
filter, no transform).
OpenAI (Responses) and Anthropic (Messages) treat xhigh and max as
DISTINCT effort levels in 2026, and the Kimi K3 wire's top tier is max —
the old parse alias (max→Xhigh) conflated them. Canonical Max now exists:
parse/as_str/serde split, Messages mapping sends xhigh and max as their
own tokens (was Xhigh→"max"), and the K3 menu token max carries
canonical Max end to end.
Kimi wire is byte-identical in all four flows (menu pick, restored
legacy xhigh session, --reasoning-effort flag, /effort command) —
adversarially traced and pinned: kimi_compat's string-level xhigh→max
rename covers legacy tokens, max passes through verbatim.
From the review:
- Rollback safety: persisted reasoning_effort (session summaries, chat
history) deserializes leniently — unknown future tokens degrade to
None with a warning instead of hiding sessions or failing resume.
- Restore migration: a pre-split xhigh override onto a model whose menu
offers max but not xhigh (K3) migrates once, healing display/active-row
drift and re-persisting the live vocabulary.
- /effort max now rejects (with the offered list) on models whose menu
lacks a max row instead of silently applying xhigh; deliberate, tested.
- The interim Responses-backend Max→xhigh downgrade (async-openai has no
Max variant through 0.41) warns loudly; real max wiring lands with the
OpenAI provider cycle via post-serialize body patch.
- Two rusted ignored-e2e wire pins asserted the pre-adapt reasoning_effort
key (deleted by the body adapter since ea0ce9d); they now pin the real
thinking.effort=max shape.
Platform API keys now live in auth.json under the platform-id scope (the
per-provider auth.json key contract), resolved env > auth.json > legacy
[platforms.*] config.toml (read-only fallback). The TUI login picker,
paste box, auth-method advertising, and authenticate handler are all
registry-generic: a new PlatformSpec row appears in the login UI and
authenticates with zero UI changes. Spec rows gained vendor/console_host/
login_label display fields (moonshot strings byte-identical, pinned by
tests).
Adversarial review caught that auth.json keys were validated at login but
never stamped onto catalog entries (completions would 401; restart lost
eager auth). Fixed red-green: resolve_model_list/resolve_model_catalog now
take a resolved PlatformApiKeys snapshot consumed by the credential-
stamping layer (auth.json beats stale config.toml, matching the login
validator), with production callers resolving fresh per catalog build.
Also from review: the new auth.json writer takes the manager's cross-
process flock (bounded retry — an unlocked RMW racing a token refresh
could revert a rotated refresh token); the oauth-401 wiremock test is
hermetic (KIGI_SHARE_DIR tempdir; it could read a dev's real auth.json
and hit live moonshot); cli_models resolves real keys; auth.json is read
once per registry sweep; caller-less lock_config_writes deleted; catalog
resolvers tightened to pub(crate); stale config.toml doc comments and the
no-credentials error copy updated.
Replace PlatformId's per-method match arms with static PlatformSpec rows
read through a single spec() accessor. Zero behavior change (adversarially
reviewed: ids, display names, base URLs + env override pairing, oauth flag,
prefix filters, and api-key env precedence order are byte-identical; all
public signatures unchanged). Groundwork for the multi-provider expansion:
adding a platform becomes variant + ALL entry + spec() arm + row.
Guards added for the row-addition workflow: spec-id uniqueness test (parse
scans rows) and a compile-forced ALL-completeness test (a variant missing
from ALL would otherwise be silently unparseable and skipped by model
sync). Future-proofed two assertions that hardcoded 'openai' as an
unknown-platform example.
README leads with the claim under the title — the world's first CLI with
built-in graph engineering — plus a dedicated section covering the /graph
command set, the closed loop it runs (plan → parallel worktree workers →
adversarial verify → 3-way merge-back → replan → optimize → terminal
whole-objective verification), the repo-following .kigi/graph.jsonl state,
copy-paste disable commands for macOS/Linux and Windows, and the tuning
knobs.
install.sh: the rc-file resolution and persist_line helper are hoisted out
of the not-on-PATH branch and generalized (guard + label), so PATH
persistence is unchanged while KIGI_GRAPH=1 is persisted unconditionally
and idempotently for zsh/bash/fish/POSIX profiles — an existing
KIGI_GRAPH line (e.g. a user's =0 opt-out) is left untouched on
reinstall. install.ps1 mirrors it with a User-scope environment variable,
also only when unset. No binary change — installers and README ship from
main, so this needs no rebuild or release.
libc::EWOULDBLOCK is unix-only (kigi-shell links libc behind cfg(unix));
the v0.1.3 windows-msvc release build failed on it. fs2 exposes
lock_contended_error() precisely as the cross-platform classifier
(EWOULDBLOCK on unix, ERROR_LOCK_VIOLATION on Windows) — use it instead
of the raw errno. No behavior change on unix; graph_project lock tests
green.
A restricted optimizer pass now reviews the graph at plan boundaries —
right after initial planning and piggybacked on each replan version
boundary, never mid-execution. An optimizer subagent may emit four ops
over Waiting/Ready nodes only: remove_dep (delete a false dependency,
restoring parallelism — the highest-value edit), reorder (pending
priority for the serial scheduler), merge (fold two tiny nodes; specs
concatenate, deps union, dependents re-point, absorbed self-deps drop),
and split (2-3 focused replacements inheriting the original's deps and
dependents). The optimizer changes graph DATA only; the executor stays
pure deterministic Rust. KIGI_GRAPH_OPTIMIZER=0 disables it entirely.
apply_optimization enforces the contract twice: per-op checks
(pending-only targets, known ids, terminal node untouchable, dead-node
deps rejected as DeadDep, merge/split targets with non-pending
dependents rejected with the true reason instead of tripping the
immutable invariant later), then FINAL invariants — every non-pending
node byte-identical in the result, the gn-final gate rebuilt over all
survivors, node cap, whole-graph acyclicity, and a BIDIRECTIONAL status
re-derivation for pending nodes (adversarial review caught the critical
hole: a merge grafting unsatisfied deps onto a Ready node would
otherwise dispatch it ahead of its new prerequisites, since
recompute_ready is promote-only). Applied passes bump plan_version,
freeze an immutable baseline, and consume a slot of the SHARED replan
cap; an explicit {"ops": []} is a respected free no-op; any failure
degrades to keeping the current plan. Plumbing reuses a new shared
artifact-pass runner (stale-artifact delete, size cap, missing-file
fail-closed) extracted from the replanner.
Tests: remove_dep parallelism restore + loud no-such-dep, immutable and
terminal-node rejections across all four ops, merge/split dependent
rewiring incl. final-gate rebuild and intra-split dep resolution,
result-cycle rejection, dead-dep splits, Ready-demote-on-merge, and
three e2e flows — false-dep removal proven ACTUALLY parallel by the
held-reply fan-out gate, OPTIMIZER=0 spawning zero passes, and the
shared-cap guard. kigi-shell 4959 lib tests green; clippy clean.
/graph show now renders the dependency graph as box-drawing text via a
deterministic Sugiyama-lite pipeline in session/graph_render.rs:
longest-path layering, dagre-style dummy pass-throughs so every drawn
edge spans exactly one layer gap (long edges route as vertical lanes
through intermediate bands), one-pass barycenter ordering, and greedy
bus-lane allocation in the connector gutters with box-drawing-aware
glyph merging (a bus crossing a pass-through renders ┼). Node boxes
carry a status glyph (✓ ▶ ○ · ✗ ⊘) and a clamped title; a legend line
closes the view. Only Blocks edges are drawn — DiscoveredFrom is audit
metadata whose origin is always terminal, so drawing it would double
edges without scheduling meaning.
The output rides ordinary scrollback (the pager already scrolls it), so
no new wire variant or pager view was needed. Honest ceiling: a graph
wider than the 120-column budget — or an empty/pre-planning graph —
degrades to the indented status tree, because box art wrapped by the
terminal is worse than no art; a cycle reaching the renderer (upstream
validation bypassed) refuses to render rather than looping.
Tests: six-node structural snapshot (layering order, fan-in bands,
arrowheads, corners, legend, no trailing whitespace, width bound),
determinism across runs, too-wide fallback, dummy-lane routing through
intermediate bands, DiscoveredFrom exclusion, title clamping, empty
graph; plus a handle_prompt e2e driving /graph show end to end (no-graph
degrade + seeded chain rendering). kigi-shell 4949 lib tests green;
workspace clippy clean.
The graph now follows the REPOSITORY, not the session. Every checkpoint
projects the orchestration to .kigi/graph.jsonl at the git root in a
beads-style, line-mergeable shape: line 1 is the header (orchestration
minus nodes — omitted entirely, and a header carrying inline nodes is
rejected on load rather than silently duplicating the per-line entries),
then one content-hash-id node per line. The session tracker stays the
single source of truth; projection failures warn loudly but never block
progress. kigi only writes the file — committing it stays a user decision.
Single-writer discipline via an fs2 flock on a sidecar .lock: the session
that creates or resumes a graph owns the projection; other instances get a
read-only /graph status view rendered from the file and an explicit
refusal on resume. Cross-session revive: /graph resume in a fresh session
loads the file UNDER the lock (locking after reading raced the owner's
final checkpoint and could resurrect a just-cleared graph), sanitizes it
with the from_snapshot demotions, and re-dispatches.
Holding the lock proves nobody writes NOW — not that the file's content is
yours. Every lock-then-mutate site therefore identity-checks the projected
graph_id: /graph <objective> refuses to overwrite a foreign non-Complete
projection (revive-or-clear guidance, mirroring the session-level guard);
session-restored graphs claim writership on resume (and best-effort at
spawn re-emit, so the shared file learns the demoted truth immediately)
but refuse when the projection belongs to a different graph; /graph clear
skips projection teardown entirely when no session graph exists, leaves
foreign projections in place, and warns instead of swallowing lock errors.
Resume arms validate their flags before taking the lock.
Tests: projection round-trip pinning the one-line-per-node shape,
inline-nodes rejection, malformed-content loud errors, exclusive-lock
semantics across handles, and an e2e driving create → checkpoint
projection → second-instance read-only refusal → owner death → fresh-
session revive (demoted node relaunches) → clear removing the projection.
kigi-shell 4941 lib tests green; workspace clippy clean.
Workers, verifiers, and serial node goals can now surface out-of-scope work
as line-anchored 'DISCOVERED: <text>' markers (fence-stripped and
placeholder-filtered — the templates' own examples are fenced so verbatim
echoes never parse; worker summaries embedded in verifier prompts get the
marker neutralized alongside NODE_RESULT/NODE_VERDICT). Discoveries queue on
the orchestration as persisted state and fold into the graph at dispatch
boundaries: a replanner subagent produces a strictly APPEND-ONLY appendix,
validated against the live graph (existing-id deps allowed; edges onto
gn-final rejected — they would cycle the moment the final-gating extension
lands; Blocks deps on Failed/Blocked nodes rejected as DeadDep so the
attempt-2 feedback loop repairs the artifact). Installing an appendix bumps
plan_version, freezes an immutable graph.baseline.v{N}.json next to the
prior versions, extends gn-final's gate (demoting a Ready final back to
Waiting), and recomputes readiness.
DiscoveredFrom edges are audit metadata, never scheduling gates: an origin
is always terminal at replan time, so gating on it is either a no-op or a
permanent wedge — and a failed node's discoveries are still real work.
Replanning is bounded by KIGI_GRAPH_REPLAN_CAP (default 3; 0 disables it
quietly): past the cap, after the final node has achieved, or on replan
failure, discoveries drain to history only — a working graph is never
paused for a failed enhancement pass, and it always converges. The budget
gate now precedes the replan boundary (a budget-dead graph keeps its
discoveries queued for a later --budget top-up instead of spending two
replanner runs first), and both planner runners delete stale artifacts
before spawning so a child that responds without writing can never get a
previous pass's file validated as its own output.
Tests: validate_replan unit coverage (existing-id resolution,
DiscoveredFrom dedup, collisions, dead deps vs dead origins, terminal-node
edges, combined-graph cycles), tracker appendix/regate/audit-edge tests,
and two e2e flows — discovery → replan → appended node runs to Achieved
with both baselines frozen, and cap-0 draining to history while the graph
still converges. kigi-shell 4935 lib tests green; workspace clippy clean.
BudgetLimited is now a resumable state: a budget trip demotes in-flight
nodes to Ready (a resource stop, not a verdict — no forever-Running node is
ever persisted) and '/graph resume --budget <tokens>' re-arms the graph with
fresh headroom (new budget = spent-so-far + extra). The tripped node's
partial burn is charged into tokens_spent_nodes at BOTH cascade sites before
the demotion clears current_node, so the top-up arithmetic never runs on an
under-counted ledger. Any input starting with 'resume' resolves to a resume
(case-insensitive; malformed top-ups surface the usage hint) and setup_graph
refuses to replace any non-Complete graph — a typo can no longer silently
destroy a resumable graph. An explicit --budget on a merely-paused graph is
rejected loudly instead of silently discarded; all trip-time messages now
advertise the top-up.
The pager gains a graph status chip: a new GraphUpdated wire variant
(extensions/notification.rs, old pagers degrade via #[serde(other)]) is
emitted from the single persist_graph_state chokepoint — every transition is
both a checkpoint and a badge tick — with a 'cleared' sentinel on /graph
clear and a one-shot re-emit after session restore (the replayed updates log
otherwise shows the pre-shutdown Active state that from_snapshot just
demoted in memory). TUI side: GraphDisplayState, session-notification arm,
and a goal-idiom chip with node progress, clamped current-node title, and
budget-aware spend. Pre-session command availability now advertises /graph
from the flags (it was fail-closed to the in-session path only, so the
welcome-screen slash menu never showed it).
Coverage: GraphUpdated wire round-trip + minimal-payload + unknown-tag
tests; PTY scenarios graph_slash_presession{,_disabled}.yaml (both run
green against the real pager binary); handle_prompt-level e2e for terminal
slash outcomes (/graph status|resume|pause, /goal refusals while the graph
owns the engine); budget top-up e2e driving a BudgetLimited diamond back to
Complete. Not shimmed: pre-G2 persisted snapshots with budget-Failed nodes
(the KIGI_GRAPH flag has never shipped enabled, so none exist).
kigi-shell 4927 and kigi-tui 6610 lib tests green; workspace clippy clean.
With KIGI_GRAPH_CONCURRENCY > 1 (default 3, clamp [1,8]) and >=2 Ready nodes,
drive_graph — the single dispatch loop shared by setup/advance/resume — runs
parallel batches: each node executes as a bounded worker<->verifier subagent
loop (KIGI_GRAPH_NODE_ROUNDS, default 3; general-purpose children with the
full implementer toolset; worktree isolation on round 1, resume keeps context
and worktree on later rounds; NODE_RESULT/NODE_VERDICT terminal contracts
parsed fail-closed with fence-stripping and line anchoring). Achieved nodes
merge back SEQUENTIALLY via kigi_workspace apply_worktree in Merge mode; a
conflict fails the node, block_dependents fires, and surviving chains keep
going. gn-final always runs serially on the full goal engine. Concurrency=1
is byte-identical to the serial G0 path; non-git projects degrade to serial.
Merge primitive hardened for real use (kigi-workspace): the 3-way apply is
now byte-safe (binary files no longer read as UTF-8 and silently deleted)
and gains the identical-content rule (ours==theirs => already present, not a
conflict) — without it, any dirty file inherited via PreserveWorkingTree
false-conflicted every node merge and multi-wave graphs self-poisoned.
Adversarial review pass (16 confirmed findings, all fixed): in-session
resume demotes orphaned Running nodes instead of wedge-pausing forever after
a mid-batch Esc; cancel re-sweeps subagents AFTER the turn abort so workers
spawned in the cancel window die too; empty child ids are never adopted as
resume targets (no unisolated escape to the shared tree); a successful
isolated round returning no worktree fails the node (soft-fallback can no
longer put N writers in one tree); main-HEAD movement during a batch aborts
merges instead of reverse-applying external commits; failed nodes still
charge the token budget; budget trips terminally fail the in-flight node;
runaway (>600s) rounds are cancelled by spawn id and retried via resume;
worker summaries are marker-sanitized before verifier embedding. Merged
worktrees are removed immediately (storage discipline); failed nodes keep
theirs for postmortem.
Tests: 4922 kigi-shell lib tests green (58 graph-specific), including
fan-out proven by a held-reply gate, real-git batch merge with cleanup
assertions, budget charging across verdicts, cap trimming, resume-after-
cancelled-batch, and backgrounded-round cancel semantics.
A deterministic DAG scheduler layered on the existing goal engine: /graph
<objective> decomposes the objective via a graph-planner subagent, gates the
result through Agentproof-style static validation (cycles, unknown deps,
duplicate slugs, caps), appends a structural final-verification node, then
executes each node as one ordinary goal — planner, worker loop, adversarial
verifier, budget and pause machinery all reused verbatim. The in-turn loop
advances nodes within the same turn (multi-loop closed loop); goal-side
auto-pauses cascade to the graph at a single chokepoint; node goals are armed
with the remaining graph budget so mid-node overruns trip graph-wide.
Gated by KIGI_GRAPH=1 (default off) + the goal harness. State persists to
<session_dir>/graph/state.json with a clear-tombstone; per-version immutable
baselines and per-node artifact archives live under graph/<graph_id>/.
Restore demotes Active->UserPaused and Running->Ready (verifier-gated re-run).
Review hardening (adversarial multi-agent pass, 24 confirmed findings fixed):
the goal-inactive loop break now consults the graph seam (mid-turn
classifier-disabled completions can no longer strand an Active graph), the
pause cascade fires even when no node goal is in flight (cancel during
planning), node bookkeeping precedes the long setup await, /graph clear only
resets the engine it owns, budget trips terminally fail the in-flight node,
and the pause transitions in /goal pause + /graph pause no longer hide inside
debug_assert! (a release-build no-op inherited from upstream).
Tests: 4908 kigi-shell lib tests green, including a serial 4-node closed-loop
e2e, restore/resume, cascade, mutual-exclusion, planning-retry, persistence
round-trip + tombstone, and status rendering.
Kigi never publishes CDN changelogs, so the entire inherited feature
was dead weight: the welcome-menu Changelog row, the hero-box info
slot (bullets + clickable CTA), /release-notes with its /changelog
alias, and the ChangelogManager CDN-fetch/disk-cache pipeline
(Effect::FetchChangelog, TaskResult::ChangelogFetched, startup and
post-login fetch kickoffs, AppView cache fields, mouse hover/click
handling). Welcome menu is now [Import] / New worktree / Resume
session / Quit; the hero box keeps title + version + subtitle and its
layout math simplifies to 3 + menu rows (verified equivalent by the
surviving boundary tests). Action::ShowReleaseNotes and the DocViewer
modal stay — /docs uses them. builtin.rs keeps deleting stale
CHANGELOG.{json,md} caches written by kigi ≤ 0.1.0.
kigi-shell-base drops its reqwest 'blocking' feature (only the deleted
module used it). -1206 lines net.
Gates: fmt clean, workspace check/clippy --all-targets 0/0, kigi-tui
lib 6609 / shell-base 56 / shell util:: 258 all passing, welcome pty
e2e (3 tests incl. braille logo) passing.
install.sh: append the export line to the login shell's rc file
(zsh: $ZDOTDIR/.zshrc; bash: ~/.bash_profile on macOS, ~/.bashrc on
Linux; fish: fish_add_path in config.fish; otherwise ~/.profile).
Idempotent — a second run detects the existing entry and skips. On an
unwritable rc it fails loudly with the manual command.
install.ps1: prepend the bin dir to the per-user PATH via
[Environment]::SetEnvironmentVariable(..., 'User') — registry-backed,
picked up by every new terminal. Strict-mode-safe for a null user Path.
Verified end-to-end on macOS with a fake HOME: install → rc appended →
fresh zsh sources it → kigi resolves and runs. All shell branches,
ZDOTDIR/XDG overrides, idempotency, and the write-failure path covered
by a harness driving the real script tail; install.ps1 parse-checked
and its path-construction logic exercised under pwsh strict mode.
The active-turn spinner next to "Waiting for response…" now cycles
the moon phases 🌑🌒🌓🌔🌕🌖🌗🌘 — the official kimi-cli's lunation
animation (rich's `moon` spinner) — matching the welcome-screen moon
logo. New glyphs::moon_spinner_frames() with the same legacy-ConHost
ASCII fallback as every other spinner set; emoji frames are 2 columns
and the status row already measures the rendered frame, so layout
adapts. The ambient "Starting session…" row and other spinners stay
braille.
Verified: kigi-tui lib 6621 + pager-render 963 tests green; real-binary
PTY e2e waiting_for_model_label_shows_before_first_token passes with
the new spinner rendering in a live vt.
Feedback about an unofficial community build belongs on its own issue
tracker, not Moonshot's feedback endpoint — and this mirrors the
official kimi-cli, whose /feedback opens its repo's issues page
(ISSUE_URL in ui/shell/slash.py). /feedback now returns
Action::OpenUrl(https://github.com/ZacharyZhang-NY/Kigi-CLI/issues),
the same battle-tested browser path /docs web uses.
The now-dead TUI text-feedback pipeline is excised: PromptInputMode::
Feedback (~ prefix composer mode), Action::{EnterFeedbackMode,
SendFeedback}, Effect::SendFeedback (the kigi/feedback ACP POST),
TaskResult::{FeedbackComplete,FeedbackFailed}, and their dispatchers.
The shell-side kigi/feedback ACP extension stays: it is protocol
surface for editor embeddings, OAuth-gated, and shared with kigi/btw.
Gates: workspace check/clippy --all-targets 0/0, fmt, kigi-tui lib
6621 passed / 0 failed.
Under the script's own Set-StrictMode -Version Latest, .Count on the
result of Where-Object throws when the filter matches nothing — which
is precisely the fresh-install case (bin dir not on PATH yet), so every
first-time Windows install ended with an error after an otherwise
successful install. Use -contains on the split arrays instead; no
member access on a possibly-null pipeline result.
Repro + fix verified under pwsh 7.5.2 with StrictMode Latest: old
expression reproduces the user's exact error, new one returns
False/True correctly for missing/present PATH entries.
The post-install guidance previously showed a one-time `export PATH=…`
that dies with the terminal. Detect $SHELL and print the persistent
command instead: append to ~/.zshrc / ~/.bash_profile (macOS) /
~/.bashrc (Linux), fish_add_path for fish (already persistent via
universal variables), ~/.profile as the POSIX fallback.
2026-07-18 11:35:39 -04:00
1622 changed files with 58886 additions and 26232 deletions
Loaded 100 of 1622 files, more files were not shown because too many files have changed in this diff.
Show more
Reference in new issue
Block a user
Blocking a user prevents them from interacting with repositories, such as opening or commenting on pull requests or issues. Learn more about blocking a user.
雷电芽衣
ZacharyZhang-NY