context-compression
Pass
Audited by Gen Agent Trust Hub on Sep 16, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONNO_CODE
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill provides instructions for the agent to summarize, archive, and compress conversation history. This creates a potential surface for indirect prompt injection, as malicious instructions embedded in previous user prompts or tool outputs could influence the summarization process or the agent's future behavior based on those summaries.
- Ingestion points: The skill ingests the entire conversation history (20+ turns) and large tool outputs (500+ lines) for compression.
- Boundary markers: The skill does not define specific delimiters or instructions to ignore embedded commands within the text being summarized.
- Capability inventory: The skill uses the Read, Write, and Grep tools as defined in the frontmatter.
- Sanitization: No explicit sanitization or filtering of the summarized content is described beyond discarding verbose details.
- [NO_CODE]: The skill consists entirely of instructional markdown and YAML configuration. It does not contain any binary scripts, scripts in interpreted languages (Python, JavaScript, etc.), or external package requirements.
Audit Metadata