github-state-manager
Track all task state via GitHub Issues and Project boards with automated column transitions and audit trail comments.
- License 许可
- GPL-3.0
- 最近更新
- 2026-04-24
- Stars
- ★ 1
- Forks
- 3
01 前言 Front Matter
- name
- github-state-manager
- next
- [ "ci-cd-pipeline", "investor-updates" ]
- group
- orchestration
- workflows
- [ "full-startup", "incident-response" ]
- description
- Track all task state via GitHub Issues and Project boards with automated column transitions and audit trail comments.
- prerequisites
- [ "plan" ]
- user-invocable
- true
03 SKILL.md 正文
GitHub State Manager
Manage task tracking through GitHub Issues and a Project board. Parse specs into Issues, move cards between columns, and post structured comments as an audit trail.
Steps
- Create a GitHub Project board with columns: Backlog, In Progress, In Review, Done.
- Parse product specs and feature checklists into individual GitHub Issues. Each Issue gets a standardized body with acceptance criteria, related file paths, and agent context.
- Apply labels for agent assignment (
website,backend,ops,research,spec) and category (feature,bug,chore,spike). - On task pickup: move card to In Progress, post a comment with agent ID and start time. Enforce single-agent assignment -- check labels before allowing pickup.
- On task completion: move card to In Review, post a comment summarizing files changed and tests written.
- On verification pass: move card to Done, post a comment with test results and deploy status.
- On context reset: read all open Issues and recent comments to reconstruct project state. Do not re-read the entire repository.
- Detect stale Issues stuck In Progress beyond a configurable threshold and flag them for review.
Anti-Patterns
- Creating Issues with no acceptance criteria, leaving agents without context to start work
- Card stays In Progress after agent finishes because no transition was triggered
- Two agents pick up the same Issue because label-based assignment was not enforced
- Context reset re-reads the full repo instead of targeting open Issues for state reconstruction
Validation
- Every spec item has a corresponding Issue with acceptance criteria and file references
- Column transitions fire on every agent state change with audit trail comments
- No two agents are assigned to the same Issue simultaneously
- Context rebuild completes in under 30 seconds using Issues, not full repo reads
04 同仓库更多 Skills
writer-memory
Agentic memory system for writers - track characters, relationships, scenes, and themes
wiki
LLM Wiki — persistent markdown knowledge base that compounds across sessions (Karpathy model)
website-creation
Build production-quality SaaS websites with opinionated design presets. Use when creating any startup website. The user MUST pick a design style before building. Enforces shadcn/ui, Figma design principles, specific CSS values per style, and anti-AI-writing. Load alongside anti-ai-writing skill. LIGHT MODE ONLY.
visual-verdict
Structured visual QA verdict for screenshot-to-reference comparisons
visual-qa-pipeline
Screenshot every page with Playwright, feed to visual QA agent for design evaluation, report results to Slack. Use after every build phase to verify UI matches the chosen design preset. Integrates with the Playwright skill and the website-creation preset system.
verify
Verify that a change really works before you claim completion