Skip to content

fix(flow): skill hints through command(); guard tests back on the PR lane - #1632

Merged
abrichr merged 1 commit into
mainfrom
claude/registry-hint-spelling
Oct 29, 2026
Merged

abrichr merged 1 commit into
mainfrom
claude/registry-hint-spelling

Conversation

@abrichr

@abrichr abrichr commented Oct 29, 2026

Copy link
Copy Markdown
Member

Main's Packages push run (40809648930) fails one test: tests/test_cli_hints.py::test_printed_hints_never_hard_code_the_engine_spelling. It flags registry.py:795-796, where #1623's pull command printed a literal openadapt-flow skill bind/qualify .... Under the launcher that names a program the person never typed. The hint now goes through cli_hints.command(), like every other hint.

Why #1623's CI passed: #1608 moved this test to nightly because it takes 6 s. That was my mistake: it's a source scan that guards every printed hint, so it belongs on the PR lane. It's back, with five safety guards the same >5 s rule moved:

  • identity: a numeric MRN homonym never verifies;
  • durable resume: never resends;
  • safety gallery: every dangerous case is safe;
  • halt detail: a typed value never appears;
  • self-repair: a healthy run never calls the agent.

Together they add about 90 test-seconds across the four PR shards.

Checked locally: test_cli_hints, the three registry test files and test_lanes pass (110 passed, 4 skipped), and ruff is clean.

…guards back on the PR lane

Main's Packages run fails tests/test_cli_hints.py::test_printed_hints_never_hard_code_the_engine_spelling
on registry.py's pull hint (#1623), which printed a literal "openadapt-flow
skill ...". Under the launcher that names a program the person never typed;
it now goes through cli_hints.command() like every other hint.

#1623's own CI never ran the test: #1608 moved it to nightly as a 6 s test,
though it is a source scan that guards every printed hint. It returns to the
pull-request lane, with five safety guards the same rule moved by mistake:
identity (a numeric MRN homonym never verifies), durable resume (never
resends), the safety gallery (every dangerous case is safe), halt detail (a
typed value never appears) and self-repair (a healthy run never calls the
agent). Together they add about 90 test-seconds across the four shards.
@abrichr
abrichr enabled auto-merge (squash) October 29, 2026 03:29
@abrichr
abrichr merged commit b604199 into main Oct 29, 2026
42 checks passed
@abrichr
abrichr deleted the claude/registry-hint-spelling branch October 29, 2026 03:41
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

1 participant