hash-calculator
Pass
Audited by Gen Agent Trust Hub on Feb 17, 2026
Risk Level: SAFE
Full Analysis
- [Prompt Injection] (SAFE): No instructions found that attempt to override agent behavior or bypass safety filters.
- [Data Exposure & Exfiltration] (SAFE): The skill performs local hashing operations; no sensitive file access or network exfiltration was found.
- [Unverifiable Dependencies & Remote Code Execution] (SAFE): The skill uses only standard library modules and has no external dependencies.
- [Indirect Prompt Injection] (SAFE): 1. Ingestion points: local files and text strings via hash_file and hash_text in SKILL.md. 2. Boundary markers: none. 3. Capability inventory: file system read and directory traversal in SKILL.md. 4. Sanitization: data is processed as raw bytes for hashing, not as instructions.
Audit Metadata