| name | walkthrough |
| description | Post-sprint narrative review. Reads PRD, tasks, git history, every changed file. Produces per-task explanations with ASCII diagrams. |
| argument-hint | [sprint e.g. v5] |
| disable-model-invocation | true |
| allowed-tools | Read, Grep, Glob, Write, Bash(git log:*), Bash(git diff:*), Bash(git show:*) |
Generate a sprint review report for sprint "$ARGUMENTS" so the developer can read and understand exactly what the code does.
Data Gathering
- Read
/Users/bv/Code/active/kapi-platform/docs/operations/sprints/$ARGUMENTS/prd.md for sprint goals and acceptance criteria
- Read
/Users/bv/Code/active/kapi-platform/docs/operations/sprints/$ARGUMENTS/tasks.md for task list and completion status
- Run
git log --oneline --all and filter for commits mentioning "sprint-$ARGUMENTS"
- For each commit, run
git show --stat {hash} to see files changed
- Read every file that was created or significantly modified
- Check
./docs/operations/sprints/$ARGUMENTS/screenshots/ for visual evidence
Output
Write to ./docs/operations/sprints/$ARGUMENTS/review.md using the format in templates/review.md.
Writing Style
- Write for a technical PM who can read code but wants the narrative first
- Lead with WHAT and WHY before HOW
- Use ASCII diagrams liberally for architecture, data flow, and component relationships
- Reference files as
path/to/file.ts:lineN so the reader can jump to source
- Be honest about what's incomplete, hacky, or needs revisiting
- Don't sugarcoat — if something is a workaround, say so
- Keep each task section self-contained (readable without reading other sections)