inspector/docs
Repository files (latest commit first)
Filename Latest commit message Latest commit date
cliffhall 313a580982 feat(skills): address Copilot review round 3 on #2204
One finding, and a sharp one: position in the stream is not causation.

`collectSkillInvocations` flattened every `tool_use` block across every
assistant event into one ordered array, and `chainHit` read position
alone. But the model can emit several `Skill` blocks in a SINGLE assistant
message, and it has not seen the first skill's body when it does — those
are parallel guesses, not a hand-off. A run that emitted `testing` and
`test-servers` together in one message scored as `A` led to `B`, and would
have gone on scoring after the pointer in `testing`'s body was deleted:
exactly the false pass this feature exists to rule out, arriving by a
second route (Copilot).

Each recorded invocation now carries the assistant event it came from —
`{payload, turn}` — and `chainHit` requires every link after the first to
land in a STRICTLY later turn. The turn count is threaded through
`collectSkillInvocations`'s `turnOffset`/`nextTurn` so a stream read in
chunks stays monotonic rather than restarting per read; without that, a
chunk boundary would reset the numbering and reintroduce the same bug on
long runs.

The scan stays greedy, which is still correct under the constraint: taking
the earliest occurrence of a link can only leave more room for the rest,
so no later starting point succeeds where the greedy one fails.

`sampleHit` is deliberately unaffected — a first-move case asks only
whether a skill fired at all, so it reads names and ignores turns.

Four new assertions in one case: a same-turn pair scores false; the same
two skills a turn apart score true; a same-turn pair does not poison a
later genuine hand-off; and only the chain's first link is unconstrained.
Plus a `collectSkillInvocations` case pinning the turn tagging and the
cross-chunk offset.

`docs/skill-authoring.md` now lists this as the second way a chained case
can false-pass, alongside a prompt carrying the target's own trigger.

Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01ABdTnqyBodw6mHVVkGehqe
Signed-off-by: cliffhall <cliff@futurescale.com>
2026-09-04 20:28:37 -04:00
..
images docs: restore the missing docs/ targets and split the catalog-seeding wording (#1811) (#1812) 2026-07-27 20:37:14 -04:00
architecture.md Merge branch 'v2/main' into v2/chore/2163-agent-skills 2026-08-29 23:06:01 -04:00
docker.md chore: extract AGENTS.md recipes into agent skills 2026-08-29 00:36:37 -04:00
inspector-roadmap-2026-h2.md chore(scripts): rename the local pre-push gate off ci, and guard it out of CI (#2146) 2026-08-26 23:56:34 -04:00
launcher-config-consolidation-plan.md docs: restore the missing docs/ targets and split the catalog-seeding wording (#1811) (#1812) 2026-07-27 20:37:14 -04:00
mcp-app-review.md fix(web): move the app origin off 6276, bound the publish route 2026-08-24 08:45:40 -04:00
mcp-server-configuration.md fix(auth): refuse a grant whose authorization server cannot be established (#2144) 2026-08-28 06:18:58 -04:00
publishing.md chore: extract AGENTS.md recipes into agent skills 2026-08-29 00:36:37 -04:00
quality-gate.md Merge branch 'v2/main' into v2/chore/2163-agent-skills 2026-08-30 01:04:58 -04:00
skill-authoring.md feat(skills): address Copilot review round 3 on #2204 2026-09-04 20:28:37 -04:00
test-servers.md chore: address Copilot review round 1 2026-08-29 00:52:45 -04:00
v1-to-v2-migration.md fix(web): bound retained app documents by size; correct the cookie claim 2026-08-24 08:45:40 -04:00