trimming-project-instructions

Pass

Audited by Gen Agent Trust Hub on Oct 30, 2026

Risk Level: SAFE
Full Analysis
  • [SAFE]: The skill contains purely instructional markdown designed to guide the agent in managing its own project instruction files (like CLAUDE.md).
  • [SAFE]: No prompt injection, data exfiltration, or obfuscation patterns were detected.
  • [SAFE]: The skill does not reference any external packages, remote scripts, or unverifiable dependencies.
  • [SAFE]: Although the skill mentions mechanical extraction using scripts, it does not provide or load external executable code, nor does it encourage unsafe execution practices.
Audit Metadata
Risk Level
SAFE
Analyzed
Oct 30, 2026, 12:20 PM