The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill processes content from a library of PDFs and documents, creating a surface where adversarial text within those documents could attempt to influence the agent's logic.
Ingestion points: Data is ingested through joelclaw docs search and API calls to docs-api:3838 as specified in SKILL.md.
Boundary markers: No explicit delimiters or boundary instructions are provided for the retrieved document content.
Capability inventory: The skill instructs the agent to derive operational principles and concrete system changes from the documents.
Sanitization: No sanitization or validation logic is defined for the incoming document text.
[COMMAND_EXECUTION]: The skill utilizes a custom command-line interface, joelclaw, to perform searches, expand context, and retrieve markdown content from the document repository.

pdf-brain