The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill is designed to analyze untrusted external data, including source code, configuration files, and security logs, which can contain malicious instructions intended to influence the agent's behavior.
Ingestion points: Instructions in SKILL.md direct the agent to map attack surfaces, review application code, and scan repositories for vulnerabilities.
Boundary markers: The skill lacks explicit instructions for the agent to use delimiters or specific ignore-rules when processing content from the repository under review.
Capability inventory: The agent has the ability to execute local scripts (scripts/scan_secrets.sh, scripts/audit_auth_surface.py) and perform network requests (scripts/check_security_headers.sh).
Sanitization: There is no evidence of automated sanitization or filtering of the ingested content before it is processed by the agent's context.
[COMMAND_EXECUTION]: The skill provides and instructs the agent to use local executable scripts for its operations.
Evidence: SKILL.md contains direct instructions: "Run the bundled scripts from the skill directory first... engineering-security-engineer/scripts/scan_secrets.sh and engineering-security-engineer/scripts/audit_auth_surface.py."
[EXTERNAL_DOWNLOADS]: The skill initiates network requests to external endpoints as part of its security validation features.
Evidence: The script scripts/check_security_headers.sh uses curl to fetch HTTP headers from a user-provided URL to perform a security posture check.

engineering-security-engineer