geo-hallucination-checker

Pass

Audited by Gen Agent Trust Hub on Mar 10, 2026

Risk Level: SAFE
Full Analysis
  • [SAFE]: No malicious patterns or behaviors were found. The skill metadata, instructions, and scripts are consistent with the stated purpose of hallucination detection. The script scripts/hallucination_checker.py uses standard libraries and processes data locally via standard input to identify suspicious linguistic patterns without external calls or persistence mechanisms.
  • [PROMPT_INJECTION]: The skill architecture is susceptible to indirect prompt injection because it processes untrusted user-provided text for analysis and rewriting. 1. Ingestion points: user drafts and external source references mentioned in SKILL.md. 2. Boundary markers: none defined in the instructions. 3. Capability inventory: no network or file-writing operations detected in the skill or script. 4. Sanitization: no explicit input sanitization. The risk is evaluated as safe due to the absence of dangerous capabilities available to the agent when using this skill.
Audit Metadata
Risk Level
SAFE
Analyzed
Mar 10, 2026, 01:58 AM