ce-compound
Pass
Audited by Gen Agent Trust Hub on Mar 15, 2026
Risk Level: SAFEPROMPT_INJECTION
Full Analysis
- [PROMPT_INJECTION]: The skill is susceptible to indirect prompt injection due to its core function of processing conversation history into documentation.
- Ingestion points: Untrusted data enters the agent context via conversation history analysis in Phase 1 (Context Analyzer, Solution Extractor).
- Boundary markers: The skill lacks explicit delimiters or instructions to ignore potential malicious prompts embedded within the conversation history being analyzed.
- Capability inventory: The orchestrator can write files to the filesystem (Phase 2) and trigger specialized subagents (Phase 3) based on the analyzed content.
- Sanitization: No evidence of sanitization, validation, or escaping of the ingested conversation data is present.
Audit Metadata