The Agent Skills Directory

[SAFE]: The skill includes explicit and robust instructions to filter secrets, credentials (API keys, tokens, passwords), and PII from external sources during the ingestion process.
[SAFE]: The included Python scripts (lint_wiki.py and lint_mermaid.py) are implemented using the standard library and are designed for read-only structural validation of the wiki files.
[SAFE]: The skill architecture establishes a clear provenance chain (Raw -> Wiki) and uses human-reviewed distillation processes to mitigate risks associated with processing untrusted external data.
[SAFE]: No malicious patterns such as prompt injection, code obfuscation, data exfiltration, or persistence mechanisms were detected in the skill's instructions or scripts.

llm-wiki