| name | releasehub-ai-toolkit |
| description | Tools and workflow helpers for the ReleaseHub repo (.ai/context/openspec). Use when analyzing a task, deciding whether OpenSpec is required, finding relevant docs/experiences, or generating a “what to read next” context set before implementing changes. |
ReleaseHub AI Toolkit
Use the bundled script to deterministically:
- Classify a task (
feature/bugfix/refactor/optimization/docs/chore)
- Extract domain + keywords
- Decide whether the task likely requires OpenSpec
- Suggest the most relevant context docs and historical experiences
Quick usage (recommended)
Run from the ReleaseHub repo (script auto-detects repo root):
python3 ~/.codex/skills/releasehub-ai-toolkit/scripts/rh_ai.py bundle --task "添加 ReleaseWindow 冻结/解冻接口,并限制冻结后不可修改配置"
Subcommands
analyze: Print task analysis JSON
experience: Search .ai/summaries/experience-index.md and print top matches
context: Suggest which docs to read next (based on analysis + keyword search)
bundle: Combined output (analyze + experience + context)
Examples:
python3 ~/.codex/skills/releasehub-ai-toolkit/scripts/rh_ai.py analyze --task "修复 XXX 接口返回不符合预期"
python3 ~/.codex/skills/releasehub-ai-toolkit/scripts/rh_ai.py experience --task "版本号格式验证边界情况" --top 5
python3 ~/.codex/skills/releasehub-ai-toolkit/scripts/rh_ai.py context --task "重构 ReleaseWindowAppService 方法命名,保持行为不变" --top 12
Notes
- Experience search reads
.ai/summaries/experience-index.md; keep it updated for better results.
- Context suggestions prioritize: task-direct docs > experiences > conventions > generic project context.