The Agent Skills Directory

[PROMPT_INJECTION]: The skill includes adversarial prompt fixtures in the tests/fixtures/ and references/ directories. These are explicitly used to verify the agent's ability to resist pressure and maintain safety boundaries (e.g., refusing to provide unverified approvals). They are part of a safety validation framework and do not pose a threat to the user.
[COMMAND_EXECUTION]: The validation script scripts/quick_validate_skill.py uses subprocess.run to execute the internal test contract tests/test_skill_contract.py. This is a standard and secure mechanism for verifying the skill's integrity and portability in a local environment.
[REMOTE_CODE_EXECUTION]: The testing utility tests/test_skill_contract.py utilizes the Python compile() function to check the syntax of the skill's own Python scripts. This is a legitimate static analysis step for automated validation and does not involve the execution of untrusted or remote code.

drift-monitor-designer