The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill processes untrusted external data to generate instructions for other agents, creating a surface for indirect prompt injection.\n- Ingestion points: Workflow Step 1 in SKILL.md instructs the agent to read user-provided paths, materials, repository context, and existing documents.\n- Boundary markers: The instructions do not specify the use of delimiters or markers to isolate external content from the agent's instructions.\n- Capability inventory: The skill is restricted to text generation and lacks capabilities for file writing, network access, or command execution in its own scripts.\n- Sanitization: The instructions do not define methods for sanitizing or filtering malicious payloads from the ingested materials.

goal-setter