The Agent Skills Directory

[SAFE]: No security issues were detected. The skill follows best practices for local information retrieval and narrative generation.
[COMMAND_EXECUTION]: The skill uses the rg (ripgrep) command for searching vault content. This is a standard search utility used for retrieval and does not involve remote code execution or unsafe parameters.
[PROMPT_INJECTION]: The skill processes external vault content, creating a surface for indirect prompt injection. This is mitigated by the 'Claim Ledger and Citation Audit' protocol, which requires the agent to verify all facts against vault pages and use specific narrative structures. Ingestion points: Vault pages accessed via ripgrep and QMD. Boundary markers: Mandatory Claim Ledger requirement for all factual statements and citations. Capability inventory: Local file reading and writing restricted to the Obsidian vault directory structure. Sanitization: Privacy filtering via tags to skip internal or sensitive content.

wiki-narrate