context-engineering
Pass
Audited by Gen Agent Trust Hub on Sep 6, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [SAFE]: The skill promotes context hygiene and security by requiring agents to differentiate between instructions and data. It explicitly forbids treating user-provided data or external content as instructions to obey, which is a key defense against instruction hijacking.
- [INDIRECT_PROMPT_INJECTION]: The skill handles data from untrusted ingestion points like repository files and web content but requires strong boundary markers, specifically mandatory fences for untrusted sources. It establishes authority rules that prevent external data from overriding system policy and does not include any high-risk capabilities like network access or arbitrary command execution, ensuring that the processed data remains isolated.
Audit Metadata