code-review
/code-review — Pre-PR Code Review
You are an AI assistant performing an automated code review of the implemented bug fix before it is submitted as a pull request. Follow the procedure below precisely.
Step 0: Load Context
- The workspace may contain multiple repositories, each with its own
agents.mdat its root. Identify the relevant repository and read itsagents.md. If it doesn't exist, stop and tell the developer to create one (point them toagents.md.template). - Read
.ai/implementation-plan.mdto understand what was planned. - Read
.ai/issue-analysis.mdto understand the original issue. - Get the full diff of all changes on the current branch vs. the base branch using
git diff. - Verify all tests pass. If tests are failing, stop and tell the developer to fix them via
/implementfirst.
Step 1: Diff Analysis
Read the full diff. For every changed line, verify it maps back to the implementation plan.
- Flag any unplanned changes (changes not described in the implementation plan).
- Verify no files were modified that aren't listed in the plan's "Files to Modify" section (unless the deviation is documented).
More from tharsanan1/wso2-se-agent-skills
create-tests
Write unit and integration tests for a reproduced bug based on issue-analysis-<issue_number>.md.
45verify-fix
Verify whether a GitHub issue is fixed in the local codebase. User provides a GitHub issue URL, the skill fetches it, extracts reproduction steps, builds the product from source, runs the reproduction steps, and reports whether the issue still exists or not.
43reproduce
Analyze a GitHub issue, reproduce the bug, and produce a structured issue analysis artifact.
42plan-fix
Plan and implement a fix for a reproduced issue using its issue analysis artifact.
18submit-fix
Create PRs for the fix across all changed repos and track everything in a local fix report.
16send-pr
Assemble and submit a pull request with proper metadata, description, and labels.
1