The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill exhibits an indirect prompt injection surface by delegating its core templates and operational semantics to an external local file.- Ingestion points: The agent is instructed to use ~/.agents/agents/SPEC.md as the 'unique source of truth' for templates and field semantics.- Boundary markers: No boundary markers or instructions are provided to the agent to ignore potential commands or malicious instructions embedded within the SPEC.md file.- Capability inventory: The skill possesses the capability to create directories and write multiple files (AGENT.json, charter.md, peers.md, STATE.md, sessions.log, AGENTS.md) to the local file system based on the patterns defined in the external file.- Sanitization: The skill does not perform sanitization or validation of the content retrieved from the SPEC.md file before using it to guide the file creation process.

agent-bootstrap