skill-grader
Pass
Audited by Gen Agent Trust Hub on Apr 2, 2026
Risk Level: SAFEPROMPT_INJECTION
Full Analysis
- [SAFE]: No malicious patterns or security risks were identified in the skill's instructions or configuration.
- [PROMPT_INJECTION]: The skill processes untrusted input from external skill files, creating a surface for indirect prompt injection. However, the risk is mitigated by the restricted toolset and the evaluative nature of the skill. (1) Ingestion points: SKILL.md Step 1 instructions specify reading the entire skill folder, including all references and scripts. (2) Boundary markers: Absent; there are no explicit instructions to ignore embedded commands in analyzed files. (3) Capability inventory: The skill is restricted to Read, Grep, and Glob tools for file system inspection. (4) Sanitization: Absent; no validation of external content is defined. Given the read-only tools and the fact that output is restricted to a grading report, the surface does not present a critical risk.
Audit Metadata