architect-troubleshoot-tool-errors
Pass
Audited by Gen Agent Trust Hub on Sep 27, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill instructs the agent to read conversation logs to diagnose tool failures, which exposes the agent to untrusted data provided by users.
- Ingestion points: The agent retrieves conversation history via the
GET /v1/convai/conversations/{conversation_id}endpoint as described in SKILL.md. - Boundary markers: The instructions do not define specific delimiters or warnings to help the agent distinguish between legitimate diagnostic information and potentially malicious instructions embedded in the user conversation.
- Capability inventory: The skill provides the agent with the ability to modify tool configurations using
PATCH /v1/convai/tools/{tool_id}. - Sanitization: There are no requirements or methods specified for sanitizing the retrieved conversation data before processing it.
- [SAFE]: The skill interacts exclusively with
api.elevenlabs.io. As the skill is authored by ElevenLabs, this represents the official vendor API and is considered safe for its intended management purpose.
Audit Metadata