managing-path-cleaning-rules
Pass
Audited by Gen Agent Trust Hub on Sep 15, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONCOMMAND_EXECUTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill instructs the agent to fetch and analyze real URL paths from the database to identify cleaning patterns. These paths are generated by external users and could theoretically contain adversarial content designed to influence the agent's regex generation.
- Ingestion points: The
execute-sqltool is used to fetchproperties.$pathnamefrom theeventstable. - Boundary markers: The instructions do not specify delimiters or "ignore instructions" wrappers for the fetched path data.
- Capability inventory: The agent has read access via
execute-sqland write access to project configuration viapath-cleaning-rules-update. - Sanitization: The workflow relies on the agent to correctly identify patterns and the user to confirm changes before application.
- [COMMAND_EXECUTION]: The skill makes extensive use of the
execute-sqltool to query analytics data and thepath-cleaning-rules-updatetool to modify project settings. These are legitimate project management capabilities used for the skill's primary purpose.
Audit Metadata