mathematical-constraint-formalization

Pass

Audited by Gen Agent Trust Hub on Feb 16, 2026

Risk Level: LOWNO_CODE
Full Analysis
  • [NO_CODE] (INFO): The skill consists entirely of markdown instructions without any associated executable scripts, binaries, or tool definitions.
  • [SAFE] (SAFE): No malicious patterns such as prompt injection, data exfiltration, or obfuscation were found in the metadata or body.
  • [PROMPT_INJECTION] (SAFE): The instructions do not contain attempts to override system prompts or bypass safety guidelines.
  • [INDIRECT_PROMPT_INJECTION] (LOW): The skill has an injection surface as it processes untrusted user-provided constraints. 1. Ingestion points: User-supplied mathematical constraints mentioned in instructions. 2. Boundary markers: None present. 3. Capability inventory: None (no subprocess, file, or network access). 4. Sanitization: None. Due to the complete lack of capabilities, this surface poses no technical risk.
Audit Metadata
Risk Level
LOW
Analyzed
Feb 16, 2026, 01:36 PM