9 factor rag
Reusable Factor Skills for Agent-native product optimization
npx -y skills add Just-Agent/just-product-factor --skill 9-factor-ragAssembled from the repository path, not quoted from the project. Check it against their README if it does not work.
One thing to look at
- 1 stars1 stars. Stars are a popularity signal and not a quality one, but at this level it is likely that nobody has read this closely except its author, and you would be relying on your own review.
SKILL.md
1.7 KB, as published. Nobody here has run it
9-factor-rag Skill
Name
9-factor-rag
Purpose
Audit and refactor RAG systems for ingestion, chunking, retrieval, reranking, grounding, citations, evaluation, and context control.
When to use
- RAG apps
- Knowledge bases
- Document QA
- Search-augmented Agents
- Citation-heavy assistants
When not to use
- Pure chatbots with no retrieval
- Static documents without query flow
Inputs
Ask for or inspect:
- README and docs
- source files
- examples and recipes
- configuration files
- tests and validation scripts
- CI/CD workflow files
- logs, changelog, release notes, or version history
Workflow
- Identify the project goal and user-facing promise.
- Inspect available files.
- Apply
checklist.md. - Score with
scoring-rubric.md. - Produce an audit report using
audit-report-template.md. - Produce a refactor plan using
refactor-plan-template.md. - Prioritize changes by product value and release readiness.
- Validate changes if the environment allows.
Review dimensions
- Data ingestion
- Chunking
- Indexing
- Retrieval
- Reranking
- Context assembly
- Citations
- Evaluation
- Failure handling
Refactor priorities
- Make data pipeline explicit
- Improve retrieval quality
- Enforce citations
- Add evaluation
- Add monitoring and fallbacks
Output format
Return:
- Summary
- Scorecard
- Critical findings
- Recommended changes
- Refactor plan
- Validation plan
- Next iteration suggestions
Example calls
See usage-examples.md.
Version
v0.5.0
Agent manifest
This Skill includes a machine-readable skill.json manifest so an Agent can quickly identify triggers, expected inputs, expected outputs, required files, and the review protocol before reading the full markdown pack.