knowledge-curator
Scans learnings and suggests topics ready for distillation into knowledge base.
- License 许可
- MIT
- 所属仓库
- goffity/dev-km
- 最近更新
- 2026-11-10
- Stars
- ★ 0
- Forks
- 0
01 前言 Front Matter
- name
- knowledge-curator
- context
- fork
- description
- Scans learnings and suggests topics ready for distillation into knowledge base.
- user-invocable
- false
02 资源文件 1
- 参考文档agents/openai.yaml187 B
03 SKILL.md 正文
Knowledge Curator
Agent สำหรับ scan learnings และแนะนำ topics ที่ควร distill เป็น knowledge base entries
Purpose
- Scan kb/05-ai-reviewed/learnings/ หา patterns
- Group related learnings by topic
- แนะนำ topics ที่มี 3+ learnings
- ช่วยเตรียม input สำหรับ
/distill - Track knowledge gaps
When to Use
- Weekly/periodic knowledge review
- เมื่อมี learnings สะสมหลายรายการ
- ก่อนใช้
/distill - เมื่อต้องการหา knowledge gaps
Instructions
Step 1: Scan Learnings Directory
find kb/05-ai-reviewed/learnings -name "*.md" -type f | head -50
find kb/05-ai-reviewed/learnings -name "*.md" | wc -l
Step 2: Extract Topics from Each Learning
For each learning file:
- Read the file content
- Extract:
- Title/slug from filename
- Tags from frontmatter
- Key concepts mentioned
- Technology/tools mentioned
Step 3: Group by Topic
Analyze and group learnings by:
- Common tags
- Similar titles
- Related technologies
- Problem domain
Step 4: Identify Distill Candidates
A topic is ready for distill when:
- 3+ learnings on same topic
- Clear pattern emerges
- Reusable across projects
- Not already in knowledge-base
Step 5: Check Existing Knowledge Base
ls -la kb/02-patterns/
Compare with candidates to avoid duplicates.
Step 6: Prioritize Topics
| Priority | Criteria |
|---|---|
| High | 5+ learnings, frequently referenced |
| Medium | 3-4 learnings, clear pattern |
| Low | 3 learnings, still evolving |
Output Format
## Knowledge Curation Report
**Learnings Scanned:** N files
**Date Range:** YYYY-MM-DD to YYYY-MM-DD
**Distill Candidates:** X topics
---
### Ready to Distill (3+ learnings)
#### 1. [Topic Name] - High Priority
**Learnings:** 5 files
**Pattern:** Brief description of the pattern
| File | Date | Key Point |
|------|------|-----------|
| `HH.MM_slug.md` | YYYY-MM-DD | Key insight |
**Suggested Distill Title:** `topic-name.md`
**Suggested Tags:** tag1, tag2, tag3
---
### Growing Topics (2 learnings)
| Topic | Count | Latest |
|-------|-------|--------|
| [topic] | 2 | YYYY-MM-DD |
---
### Knowledge Gaps
Areas with learnings but no knowledge base entry:
- [gap 1]
- [gap 2]
---
### Recommendations
1. **Distill Now:** [topic] - has 5+ learnings
2. **Watch:** [topic] - needs 1 more learning
3. **Review:** [existing entry] - might need update
04 同仓库更多 Skills
wait-what
Stop. That last message did not land: re-pitch it.
tdd
Test-driven development loop (red-green-refactor) with commit-based phase state and Jira traceability. Use when the user wants to build a feature or fix a bug test-first, mentions "tdd", "red-green-refactor", or "test-first". Replaces the tdg plugin — combines its commit/phase discipline with seam-based test quality rules from kb Testing Standards.
td
Creates post-task retrospectives with Before/After context and session documentation.
summary
Generates weekly or monthly session summaries from retrospectives and activity logs.
share
Shares knowledge across projects by copying to shared-knowledge directory with cross-project metadata.
session-analyzer
Analyzes git changes and creates retrospective drafts with Before/After context.