- Worker proxies install.sh/install.ps1 from raw.githubusercontent.com.
- dist bundle: repo, releases and issues links go to
github.com/ELY-Open-Source/Kigi-CLI, and the Gitea icon and labels become
the GitHub mark ("Star on GitHub").
Co-Authored-By: Claude <noreply@anthropic.com>
- kigi-env: update_base_url defaults to
api.github.com/repos/ELY-Open-Source/Kigi-CLI/releases.
- install.sh / install.ps1: resolve releases from the same API.
- /feedback and the no-session feedback fallback open GitHub issues/new;
/docs opens the README on GitHub.
- SECURITY, NOTICE, AGENTS, CONTRIBUTING: GitHub repo, issues and fork-based
pull requests replace git.zacharyzhang.com and AGit.
Co-Authored-By: Claude <noreply@anthropic.com>
The first release built and published on Gitea since the forge moved back.
THIRD-PARTY-NOTICES carries the new workspace crate versions; no dependency changed.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
The GitHub account is gone again, so git.zacharyzhang.com/ZacharyZhang-NY/Kigi-CLI
is the canonical repo. Reverse f7a4cb6:
- .gitea/workflows/release.yml: restore the Gitea edition retired in f7a4cb6
(draft release, per-target upload, SHA256SUMS, then publish). The Windows
leg runs on the windows-amd64 label the current runner advertises.
- .github/workflows/{release,warm-cache}.yml: drop them; they target hosted
GitHub runners and softprops, and Gitea ignores them once .gitea exists.
- kigi-env: update_base_url defaults to the Gitea Releases API again.
- install.sh / install.ps1: resolve releases from git.zacharyzhang.com.
- feedback, /docs, SECURITY.md, frontend worker + site: repo, issues and
raw-file links point at git.zacharyzhang.com.
- docs/RELEASE.md: Gitea edition of the checklist with the current runners.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
After the base URL, name, and key steps, the login now opens a model
selection screen instead of blindly saving. The listing fetch runs
concurrently with a manual-id input — a WAF challenge page or a 404 no
longer passes validation silently: a base without a /vN tail is probed
once as {base}/v1 and the working base is adopted, an anthropic-wire
listing that 401s on x-api-key retries once with Bearer (relays gate
GET /models on Bearer while /messages takes the key header), and a
failed fetch shows its reason while manual add stays usable. The chosen
ids persist as [platforms.<name>] models = [...]; catalog assembly
synthesizes selected ids the listing never served (before [model.*]
overrides so they compose) and filters unselected provider entries
(after overrides, authoritative in the provider namespace). Provider-
owned entries keep their declared wire against donor backend
inheritance, and listing errors quote line/column only — a reflected
body could carry the key.
The API-key, custom-provider, and loopback paste boxes matched
KeyCode::Char without checking modifiers, so Ctrl+V inserted a literal
'v' and no chord read the host clipboard. The arms now read the
clipboard through is_paste_key/is_inline_paste_key (the picker.rs
pattern) and the char fallthrough is gated by is_text_input_key, so
Ctrl+letter chords no longer type.
The subagent auth-scheme lookup is now a function with a test. The failed-login
span carries the same method id as the successful one. README order matches the
code: the provider is saved, then the key is checked.
The Codex backend (client_version 0.159.0) served eight list+supported models on
2026-09-29, gpt-6.1-sol first: context 272000, efforts low..max, default low.
POST /codex/responses with kigi's codex headers answers 200 and completes.
The picker lists Custom provider (OpenAI compatible) and (Anthropic compatible).
The TUI asks for a base URL, a name (default from the host), and a key, saves the
definition and the key, and the agent validates the key against GET {base}/models.
PTY e2e covers both wires against the mock inference server.
reconstruct_full_config and the subagent spawn resolved both from a disk reload
that never holds fetched entries, so a fetched Messages model (anthropic,
minimax) went out with Authorization: Bearer instead of x-api-key.
[platforms.<name>] with base_url and api (openai or anthropic) declares a provider.
Its key lives in auth.json under the name or in the table; /models is fetched per
provider, entries are keyed <name>/<model>, and the key is stamped in memory only.
The first Ctrl+C clears the draft; the second did nothing and left the user in
comment mode. Both the approval and the casual preview paths now cancel.
y copied the truncated display row. Saving or discarding a row edit left focus on
the queue pane, so the next Enter re-opened the edit instead of sending.
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.
雷电芽衣
ZacharyZhang-NYandClaude