moonlight-lupin/agent-skills / plugins/dev-pipeline/skills/review-lessons
review-lessons
VerifiedLibrary of lessons from past build-review loops — not a stage. Use when a review or fix round hits a specific case: a reviewer BLOCKING that contradicts the brief, a test that passes but checks nothing, a feature dead in production, SQLite/WAL migrations, MCP tool surfaces, REST API authz, app ↔ external-source sync, Android/APK verification, fill-template extraction, CLI/git mechanics, reviewer outages. Load the one file the case needs; never load the whole library.
references/delegate-skills-adaptation.md
View on GitHub# Adaptation Reference: delegate-skills Port (2026-09-12)
Source: https://github.com/amElnagdy/delegate-skills (amElnagdy, MIT,
reviewed Sep 2026). Verdict on the repo itself: Watch — it targets other
orchestrators (Claude Code/opencode via `npx skills`), and its relays
duplicate dispatch mechanics this skill already hardens. Concept port, not
adoption.
## Ported
1. **Dispatch result contract (`delegate-result.v1`)** — from their
`delegate-relay.result.v1`. One machine-readable result file per
dispatch: status, exitCode, model-that-actually-ran, touchedFiles,
report path. Implemented as `review-commit/scripts/capture_dispatch_result.py` +
`review-commit/references/dispatch-result-contract.md`. Their relay writes it
automatically; we capture after exit because our CLIs run raw in
background terminals.
2. **The reviewer commits** — from their core invariant "the relay never
commits — committing belongs to the reviewer" and their commit-boundary
discipline (working tree is evidence between dispatch and commit; no
reflexive `git checkout`/`clean` before inspection). Encoded as
Pitfall #32 in SKILL.md.
## Explicitly NOT adopted
- **Their relay scripts (`relay.mjs`)** — per-CLI wrappers for other
orchestrators' headless modes. Our dispatches (Cursor CLI `agent`, Codex
`exec`, Claude Code) already run with verified flags and pitfall
handling; a wrapper would add a layer without removing any known
failure mode.
- **Fleet lanes / `delegate-fleet.v1` config** — persistent lane maps with
approval-hash binding. Designed for multi-CLI fleets and multiple
orchestrator seats; this workspace runs a fixed builder/reviewer pair
with Bot Mode. Model routing lives in the brief, not in config.
- **`delegate-setup` discovery/interview flow** — multi-CLI discovery and
burn/spare economics; no fit for a 1-operator setup.
## Validation
- Script exercised 2026-09-12 on a scratch git repo: clean tree →
`touchedFiles: []`; post-edit → exact porcelain lines; non-repo dir →
`null`; exit-code mapping verified for 0/124/127/1.
Related skills
| Skill | Stars | Updated | Works with | Status | Install |
|---|---|---|---|---|---|
| finishing-a-development-branchobra/superpowersUse when implementation is complete, all tests pass, and you need to decide how to integrate the work | Stars: 323k | Updated | Claude CodeCodex | Verified | Install |
| systematic-debuggingobra/superpowersUse when encountering any bug, test failure, or unexpected behavior, before proposing fixes | Stars: 323k | Updated | Claude CodeCodex | Verified | Install |
| test-driven-developmentobra/superpowersUse when implementing any feature or bugfix, before writing implementation code | Stars: 323k | Updated | Claude CodeCodex | Verified | Install |
| skill-creatoranthropics/skillsCreate new skills, modify and improve existing skills, and measure skill performance. Use when users want to create a skill from scratch, edit, or optimize an existing skill, run evals to test a skill, benchmark skill performance with variance analysis, or optimize a skill's description for better triggering accuracy. | Stars: 198k | Updated | Claude CodeCodex | Verified | Install |