audit
Pass
Audited by Gen Agent Trust Hub on Sep 27, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONCOMMAND_EXECUTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill ingests untrusted external data from library records and CRM logs to perform strategic assessments, creating a potential path for indirect prompt injection.\n
- Ingestion points: Processes narrative library content (SKILL.md) and CRM won/lost/renewal examples (references/task-method.md).\n
- Boundary markers: The instructions lack explicit delimiters or "ignore" instructions for the data being audited, relying on the agent to distinguish between instructions and potentially malicious data.\n
- Capability inventory: The skill possesses the ability to resolve tool schemas, list agents, and perform workspace mutations or destination writes to support migrations.\n
- Sanitization: No explicit sanitization or validation logic is defined for the ingested external content before it is processed or used to drive mutations.\n- [COMMAND_EXECUTION]: The skill leverages platform-provided tools to query agent statuses, resolve tool schemas, and perform writes or migrations to the workspace and CRM entities, which are within its stated functional scope for library auditing and repair.
Audit Metadata