manage-freeze-windows
Pass
Audited by Gen Agent Trust Hub on Oct 24, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill processes data from external sources (Harness API responses via
harness_listandharness_get) that may contain untrusted content such as names, descriptions, or YAML definitions provided by other users. - Ingestion points: Data returned from
harness_listandharness_gettools as described in Step 1 and Step 2 ofSKILL.md. - Boundary markers: The instructions lack delimiters or explicit 'ignore embedded instructions' warnings when the agent processes retrieved freeze window data.
- Capability inventory: The skill possesses capabilities to modify platform state, including
harness_create,harness_update,harness_delete, andharness_execute(toggle status/manage global freeze). - Sanitization: No explicit sanitization, validation, or escaping of the retrieved content is performed before the agent acts upon it.
Audit Metadata