implement_skill
- TypeScript
0
GitHub Stars
1
Bundled Files
2 months ago
Catalog Refreshed
4 months ago
First Indexed
Readme & install
Copy the install command, review bundled files from the catalogue, and read any extended description pulled from the listing source.
Installation
Preview and clipboard use veilstrat where the catalogue uses aiagentskills.
npx veilstrat add skill matthew-plusprogramming/monorepo --skill implement- SKILL.md12.8 KB
Overview
This skill implements approved atomic specs one at a time with full traceability to requirements in an opinionated TypeScript monorepo (Express 5 + CDKTF). It executes each atomic spec, records implementation evidence, updates session checkpoints for resumption, and escalates when the spec lacks required detail. Use it after a spec group reaches APPROVED review state.
How this skill works
The skill loads the spec group manifest and atomic spec files, verifies prerequisites (approval, enforcement passing, no blocking questions), and builds an ordered implementation queue. For each atomic spec it marks status, implements code following existing patterns, runs related tests, records Implementation Evidence and Decision Log entries, updates session checkpoints, and marks the spec complete. If gaps or invalid assumptions appear, it stops, documents the issue, requests guidance, and only resumes after resolution.
When to use it
- After a spec group manifest.review_state is APPROVED and enforcement_status is passing
- When you need traceable, atomic changes tied to acceptance criteria
- When implementations must support cross-session resumption and evidence logging
- When multiple independent tasks may be parallelized safely
- Before running the verification and review chain (unify, code-review, security)
Best practices
- Always confirm manifest flags and that atomic specs exist before starting
- Follow existing codebase patterns, naming, and error handling exactly
- Update the atomic spec frontmatter and session checkpoint at each phase transition
- Record Implementation Evidence and Decision Log entries for every change
- Stop and document open questions or invalid assumptions; don’t silently deviate
- Run targeted tests after each atomic spec and full test/lint/build after all work
Example use cases
- Implement logout UI and AuthService.logout() with evidence linking components and service lines
- Execute token clearing and post-logout redirect atomic specs in order with checkpoints for resumption
- Run parallel implementers for multiple independent atomic specs while coordinating a separate test-writer subagent
- Discover a missing requirement during implementation, log an open question, get a decision, then resume
- Complete all atomic specs, run full test suite and update manifest to transition to verifying phase
FAQ
Stop implementation, document the issue as a blocking Open Question in the atomic spec, request guidance, update the spec after approval, then resume.
How do checkpoints work for interrupted sessions?
A session.json checkpoint records atomic_specs_pending and completed items; the skill reads it at start and resumes from the pending queue to avoid rework.