memory-ledger
Pass
Audited by Gen Agent Trust Hub on Sep 25, 2026
Risk Level: SAFE
Full Analysis
- [SAFE]: The skill defines a markdown-based protocol for logging task state, which is a best practice for long-running agentic workflows.
- [NO_CODE]: No scripts (Python, JavaScript, shell) or executable binaries are included in the skill. The logic is handled entirely through natural language instructions to the agent.
- [INDIRECT_PROMPT_INJECTION]: The skill is designed to ingest and store user-provided information (preferences, constraints, decisions) into a structured ledger. While this creates a surface for indirect prompt injection, the ledger is a passive documentation structure. The skill does not include any capabilities (such as shell execution or network requests) that could be leveraged for an attack based on the ingested data.
- [METADATA_POISONING]: The metadata in the YAML frontmatter and agent configuration correctly describes the skill's purpose and does not contain any deceptive instructions or hidden commands.
Audit Metadata