context-compression
Pass
Audited by Gen Agent Trust Hub on Sep 23, 2026
Risk Level: SAFENO_CODEINDIRECT_PROMPT_INJECTION
Full Analysis
- [SAFE]: The skill provides conceptual frameworks and strategies for managing large conversation contexts. It does not perform any operations or access any sensitive resources.
- [NO_CODE]: The skill does not contain any code snippets, scripts, or command executions.
- [INDIRECT_PROMPT_INJECTION]: The skill suggests strategies for summarizing conversation history. While informational, implementation of these workflows involves processing untrusted user data which is a surface for indirect prompt injection. 1. Ingestion points: Processing of conversation history in SKILL.md. 2. Boundary markers: Suggestions for structured markdown headings are provided. 3. Capability inventory: No tools or code provided in this skill. 4. Sanitization: Not addressed as this is a conceptual guide.
Audit Metadata