compile-knowledge
Pass
Audited by Gen Agent Trust Hub on Oct 23, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONCOMMAND_EXECUTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill represents an attack surface for indirect prompt injection as it instructs the agent to process session transcripts and repository content to generate durable records.
- Ingestion points: Reads from session transcripts, project files, and user-provided judgement/feedback during a task (SKILL.md).
- Boundary markers: No explicit boundary markers or instructions to ignore embedded commands within the processed data are provided.
- Capability inventory: Performs file system operations including searching (grep), editing existing files, and creating new markdown files in the
.claude/.../memory/andwiki/directories. - Sanitization: There is no mention of sanitizing or escaping the content extracted from transcripts before it is written to the persistent store.
- [COMMAND_EXECUTION]: The skill documentation references the use of external CLI tools for memory management.
- Evidence: Mentions
5dive memory addand5dive memory consolidatefor automating knowledge persistence. These are identified as vendor-provided tools belonging to '5dive-ai' and are used for their intended purpose within the skill's workflow.
Audit Metadata