get-started-with-skilder
Pass
Audited by Gen Agent Trust Hub on Oct 30, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill processes instructions and roles defined within the workspace's catalog, which could contain malicious or conflicting instructions from other sources.
- Ingestion points: Ingests external content via Step 2 (Onboarding Concierge role) and Step 3 (existing roles, skills, and tools published in the workspace).
- Boundary markers: The instructions do not explicitly specify delimiters or warnings for the agent to ignore instructions embedded within the learned roles.
- Capability inventory: The skill utilizes
init_skilder,learn, andmanagetools, which allow for reading workspace state and creating new roles or skills. - Sanitization: No explicit sanitization or validation of the fetched instructions is mentioned before they are processed by the
learntool.
Audit Metadata