Files
ai-agent/sidecar/test_skills_list.py
Gabriel Vidal a698f5b929 feat(composer): guideline chips and skill mentions follow "Runs on"
Prompt shortcuts are now shared across runners or scoped to one: `shared` +
`runner` ("local" = the lab, or a worker id). The composer resolves the
runner its prompt lands on (the pick, else the account's online worker, else
the lab) and offers the shared chips first, then that runner's; switching
runner seeds the newly revealed chips' default-on cascade. The resume box
uses the session's own runner. Settings → Prompt shortcuts becomes a Shared
section plus one tab per runner (Lab, each paired worker), with a Shared
checkbox per card and reordering kept within a section. Settings v3
migrates existing configs: worktree, commit, pr, notify, ask, clarify and
tasks become shared; everything else stays the lab's.

A worker run also lists that machine's skills instead of the lab catalog:
the sidecar's new GET /skills reports the default cwd's and the user-level
skills, the hub proxies it at /api/workers/{id}/skills (60 s cache, empty
when offline), and the composer swaps its mentions when the runner is a
worker (no lab agents there either).

Co-Authored-By: Claude Fable 5.1 <noreply@anthropic.com>
2026-10-06 16:32:47 +02:00

2.6 KiB