The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill identifies an attack surface where instructions provided in the input data could influence the agent's output, which is intended to be consumed by other 'implementer agents'.\n
Ingestion points: User interaction and the NOTES.md file (referenced in SKILL.md).\n
Boundary markers: The instructions do not specify any delimiters or warnings to ignore embedded instructions within the ingested data.\n
Capability inventory: The skill is authorized to create, edit, and delete files within the workflows/ directory and update NOTES.md (referenced in SKILL.md).\n
Sanitization: No input validation or escaping mechanisms are described to prevent malicious data from being processed.

loop-me