The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill is designed to read project configuration data from files like CONTROL_MATRIX.yaml and AI_BOM_POLICY.yaml to generate markdown documentation. This establishes an indirect prompt injection surface.\n- Ingestion points: The agent reads .agile-v/CONTROL_MATRIX.yaml, config/control_matrix.yaml, and templates/AI_BOM_POLICY.yaml as sources for document generation.\n- Boundary markers: There are no explicit instructions or delimiters provided to ensure the agent ignores or sanitizes potential instructions contained within these configuration files.\n- Capability inventory: The agent has the capability to write multiple files across a structured docs/ directory, which could be used to output injected content or exfiltrate data through the repository.\n- Sanitization: The skill contains specific instructions to redact personal data and secrets, which mitigates direct exposure but does not fully address the risk of the agent being manipulated by malicious content within the ingested YAML files.

documentation-agent