| name | grillmester-review |
| description | Self-review the complete task-scoped diff before independent review or delivery. Use after implementation, before a pull request, or when asked to inspect uncommitted work, a branch against a fixed point, or changes since a named revision. |
Review the complete task diff
Establish the boundary
Read the acceptance criteria, locked decisions, and repository instructions
that govern the task. Resolve the stated base or fixed point, then account for
staged, unstaged, committed, and untracked changes in scope. Read every new
untracked file in full because ordinary diff output omits it.
Stop when the diff is empty, the base is ambiguous, or unrelated work cannot
be separated safely. Review primary evidence, not the implementation summary or
memory of the intended change.
Review six axes separately
- Correctness — trace changed control and data flow; check errors, state
transitions, cleanup, concurrency, and failure handling.
- Regression — find affected callers, contracts, shared defaults,
migrations, and behavior outside the immediate edit.
- Edge cases — test relevant empty, missing, malformed, repeated,
concurrent, timeout, retry, and partial-failure cases.
- Requirement coverage — map every acceptance criterion and locked
decision to concrete code or test evidence.
- Repository standards — check established architecture, naming,
language, test, documentation, and delivery rules.
- Scope — require every hunk to serve the task and identify requested
behavior the diff does not implement.
When the change touches personal or confidential data, authentication or
authorization, secrets, logging, network trust, external integrations,
deployment permissions, or another security boundary, invoke and apply
grillmester-security-review.
Correct and verify
If the active caller is the writer, fix in-scope findings before returning. When
reviewing delegated work, do not edit the implementation in the review
context; return the smallest bounded correction to the active writer instead.
Discover deterministic gates from repository instructions, build files, test
configuration, and CI rather than assuming a toolchain. Run the smallest
relevant checks and required final gates after the last edit. Record commands,
relevant results, and exit codes; distinguish fresh evidence from stale or
unverified claims.
Report material findings first, ordered by consequence. For each finding, give
the affected path and line when available, the concrete failure mode, and the
smallest useful correction. End with acceptance-criterion coverage, gate
evidence, and remaining uncertainty. A clean review must still state what was
inspected and what the evidence does not prove.