skills/q00/openclip/oc-cut-judge/Gen Agent Trust Hub

oc-cut-judge

Pass

Audited by Gen Agent Trust Hub on Sep 5, 2026

Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
  • [INDIRECT_PROMPT_INJECTION]: The skill reconciles "cut proposals" from external sources, which are interpolated into the prompt. It does not utilize specific boundary markers or instructions to prevent the model from following commands that might be embedded within these external inputs.
  • Ingestion points: Ingests transcript section proposals from other agents or files (SKILL.md).
  • Boundary markers: Absent. There are no delimiters or "ignore instructions" warnings for the processed data.
  • Capability inventory: The skill uses Bash, Read, and Write tools to manage directories and output JSON files (SKILL.md).
  • Sanitization: Absent. Data is reconciled based on timestamps without filtering for non-data content.
Audit Metadata
Risk Level
SAFE
Analyzed
Sep 5, 2026, 10:34 PM
Security Audit — agent-trust-hub — oc-cut-judge