jairorodriguezarias/siesta/kb-manager
Claude Code and agent skill by jairorodriguezarias/siesta
- community
Checking that the SKILL.md file is still in the repository.
Manages the JSON-based knowledge graph with progressive disclosure. Use when reading or writing decisions, blockers, learnings, or consultations to the KB. Provides query, append, and link operations.
Install
npx vskill@latest install jairorodriguezarias/siesta/kb-managerSKILL.md
KB Manager
Manages the file-based JSON knowledge graph. Nodes store typed information (decisions, blockers, learnings, consultations) with summary + detail for progressive disclosure. Edges store typed relationships between nodes.
Pipeline Tool Boundaries
When tools are disabled (consultant, proxy or learner), use the KB context provided in the prompt and return the requested text protocol. The CLI examples below are reference operations for an operator or a tool-enabled role; do not narrate their execution. Python persists protocol decisions, completion records and parsed learnings. A worker must never create an "Issue #N completed" node or alter checkpoints to bypass Python's test gate.
When to Use
- Before executing an issue: query relevant KB context
- After making a decision: log it to KB
- After resolving a blocker: log resolution to KB
- After learning something: log it to KB
- NOT for: trivial changes that don't affect future work
KB Structure
Progressive Disclosure
Level 1 — Summary only (default, cheap):
Level 2 — Full node (on demand, when an issue needs detail):
Level 3 — Filtered by type (when you need all decisions, or all blockers):
Operations
Query (read)
Append Node (write)
Append Edge (link)
Initialize (new project)
Common Rationalizations
| Rationalization | Reality | |
|---|---|---|
| "I'll remember this decision" | You won't. The KB is the memory. Log it. | |
| "Detail is too long to write" | Summary is enough for Level 1. Detail can be long — it loads on demand. | |
| "I don't need to query the KB, I know the context" | Past decisions affect current work. Always query summaries first. | |
| "This is too small to log" | If a future issue could benefit, log it. Small decisions compound. |
Red Flags
- Writing detail without a summary (summary is the progressive disclosure entry point)
- Creating nodes without edges (orphan nodes are hard to discover)
- Loading full detail when summaries would suffice (wastes context tokens)
- Not initializing a fresh KB per project (cross-project contamination)
Verification
- [ ] Every decision has both summary and detail
- [ ] Blockers have resolution edges to what resolved them
- [ ] Consultations are linked to the issue they were for
- [ ] Summaries are one line, under 100 characters
- [ ] Queries start with --summary-only before drilling into detail
Security
- Security scanPassed, 95 of 100Pattern checks on the skill's files.
- Model reviewPassed
Quality evaluation
- Task evaluationNot runNo evaluation yet. Run one in Skill Studio.
Source
- SKILL.md in the repositoryCheckingfactory/skills/kb-manager/SKILL.md
Popularity
- Repository stars
- 79
- Repository forks
- 18
- 7-day trend score
- 22
Stars and forks are for the whole GitHub repository, not this skill alone. Not refreshed yet.
Works with
Any agent vskill installs to: 56 targets in all. These 39 have a page on this site.
- Claude Code
- Cursor
- GitHub Copilot
- Windsurf
- Codex
- Gemini CLI
- Cline
- Amp
- Roo Code
- Goose
- Continue
- Kilo Code
- Junie
- Trae
- OpenHands
- Qwen Code
- AI Agent
- Tabnine Chat
- Sourcegraph Cody
- Amazon Q
- Firebase Genkit
- Sweep
- Devin
- Phind
- Mentat
- Claude Desktop
- Aider
- GPT Pilot
- Plandex
- bolt.new
- v0
- Replit Agent
- Anse
- TypingMind
- BetterChatGPT
- LibreChat
- Warp AI
- Pieces
- Fig AI
Embed badge
[](https://verified-skill.com/skills/jairorodriguezarias/siesta/kb-manager)