nntrivi2001
f887c16309
Pricing: the LiteLLM sync now adds every current Claude / GPT (>= 5) /
Gemini (>= 3) chat model before it is used and keeps its own rows
(model_pricing.source = 'sync', the Auto badge) at the published rate. A
user's row (any PUT) and a deleted pattern (model_pricing_dismissed) are
never touched; a deleted built-in row is not re-seeded at the next start; a
reset re-applies the list at once; retired Claude models are priced only
once used.
Models: GET /api/ai-tools/:id/models reads each tool's own list (Codex's
models_cache.json, Claude's aliases named after the newest version plus
priced versions, a catalog tool's vendor rows plus the models its sessions
used). A catalog `model: {flag, vendor?}` field passes `<flag> <id>`
(validated: never a flag or shell text); a per-tool default model starts
each run, and a resume keeps the session's own model.
Settings: every installed, enabled catalog tool without its own tab gets
one (AiToolConfig.tsx): path, homepage, Update, sessions/cost/lanes,
default model, vendor pricing, and the config files its release catalog
lists, edited over the opened sha, syntax-checked, backed up, written
atomically; credential-looking names refused. Antigravity keeps its hooks
card there.
Lanes: a just-started run no longer reads "dead" (age counts the run's
start while the old session is still bound); a second dashboard on the
same database re-broadcasts the lanes the other changed (PRAGMA
data_version, every 2 s); a lane start with a bad model answers 400.
DB: every migration runs under a cross-process lock (<db>-migrate.lock), so
two dashboards started at once on a fresh database both come up.
AI Tools catalog
One JSON file per AI coding CLI AgentHub can run in a lane. Loaded by
server/lib/ai-tools.js; the fields are described in
docs/superpowers/specs/2026-09-28-ai-tools-a1-design.md § 1. A file in
DASHBOARD_AI_TOOLS_DIR with the same id replaces the built-in one (it may
not set launcher).
Install / update (A2)
Two optional fields, honoured only in built-in files and in the copy of the
newest release's catalog that server/lib/ai-tools-sync.js keeps in
<data dir>/runtime/ai-tools-remote/ — a file in DASHBOARD_AI_TOOLS_DIR
carrying either is rejected, because an Install click would run what it names:
install:{ "kind": "npm", "package": "<npm name>" }(runsnpm install -g <package>@latestwith the login shell's npm) or{ "kind": "script", "url": "https://…" }(https, same site — last two host labels — ashomepageUrl; downloaded by the server, run asbash <file>).update:{ "kind": "npm" }or{ "kind": "script" }(reuseinstall), or{ "kind": "self", "args": ["update"] }(the tool's own updater).
Before adding an npm package, check its owner (npm view <pkg> maintainers repository.url) and that its bin matches detect: a squatted name would
install a stranger's code on every machine that clicks Install.
Attribution
The commands, prompt modes, permission-bypass flags and homepages of the tools
other than Claude Code, Codex and Antigravity come from Orca
(https://github.com/stablyai/orca, MIT License, Copyright (c) Stably AI):
src/renderer/src/lib/agent-catalog.tsx, src/shared/tui-agent-config.ts and
src/shared/tui-agent-permissions.ts.