tk-grooming
Pass
Audited by Gen Agent Trust Hub on Sep 3, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill processes potentially untrusted content from existing skills, rules, and auto-memory, creating a surface for indirect prompt injection that could influence the auditor's findings or modification proposals.\n
- Ingestion points: Workflow Step 2 in
SKILL.mdand the Evidence Chain inreferences/persistent-context.mddescribe reading native skill, rule, and auto-memory paths.\n - Boundary markers: While the skill preserves managed markings, it does not implement explicit delimiters to separate the content of audited files from its own operating instructions.\n
- Capability inventory: The skill is capable of performing file system operations including
move,tighten,delete, andfixon skill and instruction files.\n - Sanitization: The skill mitigates risks through mandatory checkpoints (Step 8), item-level approval requirements for semantic changes, and host-native question surfaces for user validation.
Audit Metadata