anth-enterprise-rbac
Pass
Audited by Gen Agent Trust Hub on Sep 18, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONEXTERNAL_DOWNLOADS
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill contains an attack surface for indirect prompt injection as it processes external organization data and role mappings.\n
- Ingestion points: Workspace maps, role matrices, and service inventory lists (SKILL.md).\n
- Boundary markers: Instructions explicitly mandate redacting credentials, prompts, and personal data from output receipts to prevent leakage.\n
- Capability inventory: Python code templates utilize the
anthropicSDK for model interaction; the skill is authorized to useWriteandEdittools.\n - Sanitization: The instructions require enforcing application-level RBAC constraints on models, tokens, and rate limits before message construction, reducing the risk of unauthorized model usage.\n- [EXTERNAL_DOWNLOADS]: The skill references the
anthropicPython package as a dependency for its implementation logic. This is a legitimate package from a well-known service provider required for the skill's functionality.
Audit Metadata