The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill is designed to process external research data from the web (e.g., census records, passenger manifests, and burial logs) and incorporate it into a local markdown-based vault.
Ingestion points: The skill ingests data from web search results retrieved during the execution of autonomous research prompts (e.g., prompts/03-findagrave-sweep.md and prompts/11-immigration-search.md) and through the manual seeding of family tree files.
Boundary markers: The prompt templates include specific ## Guard Rails and ## Protocol sections designed to define successful outcomes and prevent the agent from accepting unverified or hallucinated sources.
Capability inventory: The skill facilitates the agent's use of network search tools to gather information and file system access to create or update markdown files in the user's research vault.
Sanitization: The skill employs structured markdown templates and YAML frontmatter to organize data, but it lacks programmatic sanitization, relying instead on the agent's adherence to the provided research methodology and manual human review checkpoints.

autoresearch-genealogy