data-loss-gate
Pass
Audited by Gen Agent Trust Hub on Oct 7, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [SAFE]: The skill is a safety-enhancing utility designed to protect user data by requiring explicit confirmation for destructive operations.\n- [INDIRECT_PROMPT_INJECTION]: The skill exhibits an indirect prompt injection surface through its inspection of external metadata and file content used to populate confirmation cards.\n
- Ingestion points: The skill reads untrusted data using
get_backlinks,gbrain graph,git log, and target file contents as specified inSKILL.md(Step 1).\n - Boundary markers: Data is presented within a structured 'Confirmation Card' markdown template in
SKILL.md(Step 2).\n - Capability inventory: The skill gates destructive shell commands and gbrain CLI operations identified in the 'When This Fires' section of
SKILL.md.\n - Sanitization: The instructions do not define specific sanitization or escaping mechanisms for the metadata ingested from file headers or git logs before display.
Audit Metadata