architect-manage-procedures
Pass
Audited by Gen Agent Trust Hub on Sep 27, 2026
Risk Level: SAFE
Full Analysis
- [SAFE]: The skill documents standard REST API interactions for agent procedure management. All documented operations target the official ElevenLabs API domain (api.elevenlabs.io), which is consistent with the skill author identity.
- [CREDENTIALS_UNSAFE]: No hardcoded credentials were found. The skill appropriately uses variable placeholders ($API_KEY, $AGENT_ID, $BRANCH_ID) for sensitive identifiers and authentication tokens.
- [INDIRECT_PROMPT_INJECTION]: The skill establishes a process for reading procedure content from an external API (Ingestion: GET /v1/convai/agents/$AGENT_ID/branches/$BRANCH_ID/procedures/$PROCEDURE_ID). While this introduces an attack surface, the skill is scoped to vendor-owned infrastructure and includes guidance on using proper JSON serialization to prevent common formatting vulnerabilities. Capability inventory: restricted to standard API read/write operations for procedure management. Sanitization: use of JSON serializers is explicitly recommended for handling multi-line content. Boundary markers: not explicitly defined in the provided file. Capability inventory: restricted to authenticated REST API calls.
Audit Metadata