development-steward
Run the scheduled, read-only weekly maintenance review for the five primary product repositories and prepare one concise private report for Check-in.
- License 许可
- MIT
- 最近更新
- 2026-11-14
- Stars
- ★ 0
- Forks
- 0
01 前言 Front Matter
- name
- development-steward
- description
- Run the scheduled, read-only weekly maintenance review for the five primary product repositories and prepare one concise private report for Check-in.
02 资源文件 3
- 参考文档agents/openai.yaml248 B
- 参考文档references/prompt.md1.7 KB
- 脚本scripts/runner.mjs5.8 KB
03 SKILL.md 正文
Development Steward
Run headlessly every week. The initial allowlist is exactly AO HyS, Casa Roca, The Barber Central, NutriPlan, and ETERIA. Never add another repository without explicit opt-in.
On macOS, install the real Monday 09:00 local schedule with development-steward-schedule-enable, passing the absolute projects root and Codex executable. Confirm it with development-steward-schedule-audit; use development-steward-schedule-disable to unload it without deleting prior reports. The managed private report is ~/.development-system/steward/reports/latest.json. The LaunchAgent runs ephemeral Codex in a read-only sandbox and does not use a shell. Its raw evidence must pass the installed deterministic Steward core and Check-in core before replacing the last complete report.
Inspect skills, Codex Security, React, TanStack, shadcn, Convex, Cloudflare, Expo/mobile, PostHog, Release Train, and repository-declared fitness functions. Compare pinned upstream versions using both an exact diff and changelog. Never adopt latest, and never use stars as the only adoption signal.
Discovery is read-only. Keep repository failures local and report missing or stale evidence as unproven. Produce one short private report with what changed, what remains healthy, what needs action, links/evidence, and whether each human action is mobile or computer work. Feed that report into Check-in.
A deterministic safe update may prepare a branch and draft PR after focused checks. It must never auto-merge, release, promote production, run destructive migrations, or delete tracker state. Those operations keep their separate human authorization.
04 同仓库更多 Skills
working-backwards
Define a software feature through PRFAQ and staged documents when the user requests working backwards before coding.
flow-implement
Implement a requested software change through verification, corrections, and its authorized delivery endpoint.
development-steward
Run the scheduled, read-only weekly maintenance review for the six explicitly opted-in repositories and prepare one concise private report for Check-in.
grill-with-docs
Run one recoverable interview for functional, visual, or mixed software grills; reuse equivalent answers and offer the shared HTML questionnaire.
global-agent-guardrails
Install, audit, test, or update the Development System's global destructive-command guard for Codex, Claude Code and T3 Code. Use when configuring agent safety hooks or investigating whether dangerous shell operations are blocked.
impeccable
Design or refine frontend UI using project context and a focused design playbook; also handles visual critique and polish.