loop-me
Pass
Audited by Gen Agent Trust Hub on Oct 26, 2026
Risk Level: SAFE
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill ingests data from
NOTES.mdandworkflows/*.mdto inform its workflow design process, which constitutes an attack surface for data-driven influence on the agent's logic. This behavior is fundamental to the skill's purpose as a workflow specification tool. - Ingestion points: Reads from
NOTES.mdand various markdown files in theworkflows/directory. - Boundary markers: No explicit boundary markers or "ignore instructions" delimiters are defined for processed files.
- Capability inventory: The agent is instructed to create, edit, and delete files within the
workflows/directory. - Sanitization: No explicit validation or sanitization logic is specified for the content ingested from the workspace files.
- [COMMAND_EXECUTION]: The skill's core logic involves managing a set of workflow specification files. It directs the agent to perform file system operations, including creation, modification, and deletion of markdown files. These operations are scoped to the defined workspace directory (
workflows/) and are required for the skill's primary function.
Audit Metadata