The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill is designed to ingest and process untrusted data describing external applications and architectures as part of its security audit function. While this constitutes an attack surface where malicious instructions could be embedded in the audit inputs, the skill lacks the capabilities required to facilitate a high-impact exploit. * Ingestion points: application_description, architecture_details, deployment_context, data_sensitivity, existing_controls, specific_concerns, and testing_authorization fields defined in SKILL.md. * Boundary markers: The skill relies on natural language instructions for safe testing but does not use explicit technical delimiters. * Capability inventory: The skill is limited to text generation; it does not request or use shell execution, network access, or file system write capabilities. * Sanitization: No explicit sanitization or validation logic is defined for the input fields.

owasp-llm-top10