autopilot-status
Pass
Audited by Gen Agent Trust Hub on May 15, 2026
Risk Level: SAFECOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill executes a local Python script named
campaign-health-monitor.pywith various subcommands (health-score,corrections-history,savings-report) to process campaign data and generate reports. - [INDIRECT_PROMPT_INJECTION]: The skill ingests data from external sources including brand profiles and agency SOPs. This creates a potential surface where malicious instructions could be embedded in these data files to influence agent behavior.
- Ingestion points: The skill reads brand configuration from
~/.claude-marketing/brands/{slug}/profile.jsonand agency procedures from~/.claude-marketing/sops/. - Boundary markers: No explicit boundary markers or XML tags are used to delimit external data from the system prompt.
- Capability inventory: The skill has the capability to execute shell commands via the
campaign-health-monitor.pyscript. - Sanitization: There is no evidence of sanitization or validation of the content read from the JSON or SOP files before it is processed.
Audit Metadata