The Agent Skills Directory

[COMMAND_EXECUTION]: The skill requires the agent to execute shell commands (git rev-parse --show-toplevel) to resolve the project root and locate scripts. It also executes local Python scripts (rubric-score.py and validate-audit-artifact.py) to perform scoring calculations and validate generated reports.- [PROMPT_INJECTION]: The skill processes untrusted external data, including backlink exports, search visibility metrics, and entity evidence, creating a surface for indirect prompt injection. While it employs a validation script before persisting artifacts, the ingestion of raw third-party data remains a potential vector for malicious instructions embedded in metadata or content profiles.- [COMMAND_EXECUTION]: The skill instructions specify the interpolation of variables like $AARON_SKILLS_ROOT into shell commands, which could be exploited if the environment variable resolution is manipulated, though this risk is mitigated by the skill's specific file-path constraints.

domain-authority-auditor