token-invariants

Pass

Audited by Gen Agent Trust Hub on Mar 27, 2026

Risk Level: SAFENO_CODE
Full Analysis
  • [SAFE]: The skill consists entirely of natural language instructions guiding an AI agent on how to perform token invariant reviews. It contains no executable scripts, shell commands, or binary files.
  • [SAFE]: No sensitive data exposure, credential usage, or network-based exfiltration patterns were detected within the documentation or metadata.
  • [SAFE]: The skill does not attempt to override agent safety constraints or perform any form of prompt injection.
  • [NO_CODE]: There are no software dependencies, package manifests, or remote code download patterns associated with this skill.
Audit Metadata
Risk Level
SAFE
Analyzed
Mar 27, 2026, 06:43 AM