| name | reviewer:result |
| description | Review implementation against a spec using git diffs to catch mismatches, missing work, and bugs. Supports iterative review+fix with --fix flag and parallel multi-angle review. Use after implementing a feature to verify spec coverage. |
Result Review
Review implementation against a spec by analyzing git diffs and cross-referencing with spec requirements.
Usage
Provide spec file paths and optionally a base branch:
$result docs/plans/auth-flow/
$result spec.md tasks.md --base develop
$result docs/plan/ -b main
With iterative auto-fix:
$result docs/plan/ --fix
$result docs/plan/ --fix --base main -n 5 --fix-all
$result docs/plan/ --fix --parallel "security,coverage"
If no paths are given, ask which spec files to review against.
Process
- Read the spec files provided
- Dispatch the result-reviewer agent (or result-fix-orchestrator with
--fix)
- Report findings back
Agent Dispatch
Without --fix: use the companion result-reviewer.md in this directory as the agent prompt, and provide log_script_path — resolve ${CLAUDE_PLUGIN_ROOT}/scripts/log-findings.sh to an absolute path.
With --fix: dispatch reviewer:result-fix-orchestrator with parameters (spec_paths, working_directory, base_branch, max_iterations, parallel, angles, fix_all, review-angles template content, log_script_path — resolve ${CLAUDE_PLUGIN_ROOT}/scripts/log-findings.sh to an absolute path).
The agent will cd to the git root automatically. Provide it with:
- The list of spec files/folders to review against
- The base branch for diff comparison (or "auto-detect")
