compact-brief
Pass
Audited by Gen Agent Trust Hub on Oct 6, 2026
Risk Level: SAFE
Full Analysis
- [SAFE]: The skill is a utility for context management and does not exhibit malicious behavior. It includes security-conscious instructions such as excluding sensitive data like secrets or tokens from generated summaries.
- [INDIRECT_PROMPT_INJECTION]: While the skill ingests external data via arguments and file paths, it incorporates explicit defensive boundaries by instructing the agent to treat carried-forward content as plain data to be quoted, rather than as executable instructions.
Audit Metadata