Rag Pipelines
v1.0.0Deep RAG workflow—document ingestion, chunking, metadata, retrieval and reranking, grounding and citations, evaluation, and failure modes (hallucination, sta...
⭐ 0· 92·0 current·0 all-time
byClawKK@codekungfu
MIT-0
Download zip
LicenseMIT-0 · Free to use, modify, and redistribute. No attribution required.
Security Scan
OpenClaw
Benign
high confidencePurpose & Capability
The name/description (deep RAG pipeline) matches the content of SKILL.md. There are no unrelated environment variables, binaries, or install steps required — everything requested is proportional to documenting a workflow.
Instruction Scope
SKILL.md contains guidance and checklists for ingestion, chunking, retrieval, grounding, and evaluation. It does not instruct the agent to read arbitrary files, exfiltrate data, access unspecified env vars, or contact external endpoints. Instructions are scoped to design and debugging of RAG systems.
Install Mechanism
No install specification or code files are present (instruction-only). This is low-risk: nothing is downloaded or written to disk by the skill itself.
Credentials
The skill declares no required environment variables, credentials, or config paths. It does reference other skills (vector-databases, llm-evaluation) which may have their own requirements, but this skill itself does not request secrets.
Persistence & Privilege
always is false and the skill is user-invocable; it does not request persistent system presence or modification of other skills' configurations.
Assessment
This skill is a benign, instruction-only checklist for building and debugging RAG systems. It does not itself install code or ask for credentials, so the immediate risk is low. Before enabling it widely, consider: (1) other skills it references (vector-databases, llm-evaluation) may require API keys or installs—review those separately; (2) the agent running this guidance may have system/network access—ensure the agent's permissions are appropriate, since following the guidance could lead to actions (ingesting documents, connecting to databases) that do require credentials; (3) because it’s instruction-only, behavior depends entirely on the agent and other connected tools—audit those integrations if you plan to run the workflow in production.Like a lobster shell, security has layers — review code before you run it.
latestvk97bp7v88js4pys85knt3xv1pn83jq2g
License
MIT-0
Free to use, modify, and redistribute. No attribution required.