omh-memory-sync
Pass
Audited by Gen Agent Trust Hub on Sep 25, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONCOMMAND_EXECUTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill processes content from USER.md, MEMORY.md, and external provider context which could contain untrusted data.
- Ingestion points: Data is read from USER.md and MEMORY.md during the 'Claim extraction' and 'Inventory' phases in SKILL.md.
- Boundary markers: The skill explicitly labels external content as 'not_omh_reviewed' and uses a 'Per-entry confirmation' protocol (SKILL.md) to quote claims back to the user.
- Capability inventory: The skill utilizes the 'omh' CLI tool suite (SKILL.md and references/source-recovery.md) for memory status, recall, attention management, and pruning.
- Sanitization: Instructions in SKILL.md require quoting claims exactly and obtaining individual user confirmation for each revision or archive action before a diff is prepared.
- [COMMAND_EXECUTION]: The skill relies on various omh CLI commands to perform its core memory management functions.
- Evidence: SKILL.md and references/source-recovery.md reference multiple omh memory sub-commands, including status, recall-incident, sources, attention, confirm, retire, and prune for managing record lifecycles.
Audit Metadata