The Agent Skills Directory

[COMMAND_EXECUTION]: The skill uses the local Python interpreter to run internal validation and utility scripts, such as scripts/secfind.py for ID generation and SARIF export. It also references an environment-level locking utility (lock_scan.py) to prevent concurrent scan conflicts. These executions are part of the core functionality and do not involve untrusted remote code.
[INDIRECT_PROMPT_INJECTION]: By design, the skill ingests and analyzes untrusted source code from repositories. This creates an attack surface where a malicious repository could contain instructions intended to influence the scanning agent. However, the skill mitigates this by using structured JSON schemas for findings and coverage, and by operating under a strict security contract that emphasizes evidenced-based reporting over arbitrary instruction following.
[SAFE]: The skill includes proactive security controls, such as regex-based detection in scripts/seclib.py to identify and mask sensitive credentials (AWS keys, GitHub tokens, private keys) before they are recorded in logs or reports. The documentation explicitly forbids network scans, unauthorized repository access, and the inclusion of secrets in output, aligning with security best practices.

sec-scan