The Agent Skills Directory

[COMMAND_EXECUTION]: The script scripts/script_tester.py uses subprocess.run to execute Python scripts found within a user-provided skill directory. This behavior is the primary intended function of the 'Script Tester' tool, used to verify runtime functionality and output formats of the scripts under test. The execution is performed using shell=False to mitigate shell injection risks.
[DYNAMIC_EXECUTION]: scripts/script_tester.py facilitates the runtime execution of local Python files discovered in the target directory. Additionally, references/quality-scoring-rubric.md contains literal strings of eval() and exec() calls within a 'Security Best Practices' section. These are documented anti-patterns provided for educational purposes and do not represent executable code within the skill itself.
[INDIRECT_PROMPT_INJECTION]: The skill's primary function is to ingest and process external data (documentation and code) from other skills. This creates an attack surface where a malicious skill being audited could attempt to influence the agent's behavior through instructions embedded in its SKILL.md or code comments. The tools attempt to mitigate this by providing objective scoring rather than open-ended text processing.
[CREDENTIALS_UNSAFE]: The file tests/test_security_scorer.py contains hardcoded mock credential strings (e.g., 'super_secret_password_123', 'sk-1234567890abcdef'). These are functional test cases used exclusively to verify that the security_scorer.py regex engine correctly detects hardcoded secrets and are not active or sensitive credentials.

skill-tester