check-originality

Pass

Audited by Gen Agent Trust Hub on Jul 7, 2026

Risk Level: SAFE
Full Analysis
  • [SAFE]: No malicious code or suspicious patterns were found in the skill's instructions or scripts. The skill performs legitimate text analysis for research integrity.
  • [PROMPT_INJECTION]: The instructions in SKILL.md are focused on task execution and do not contain any attempts to override safety guidelines, ignore previous instructions, or extract system prompts.
  • [COMMAND_EXECUTION]: The skill uses a locally provided Python script (scripts/overlap_check.py) to perform text comparison. The script is deterministic, uses only the Python standard library, and does not invoke arbitrary shell commands or external processes.
  • [DATA_EXFILTRATION]: No network exfiltration or unauthorized file access was detected. The script reads only the files explicitly provided as arguments and outputs analysis results to standard output for the agent to use.
  • [REMOTE_CODE_EXECUTION]: There are no patterns of downloading and executing remote scripts, packages, or dynamically generated code. The skill relies entirely on its included scripts and standard environment tools.
Audit Metadata
Risk Level
SAFE
Analyzed
Jul 7, 2026, 01:03 PM
Security Audit — agent-trust-hub — check-originality