llm-prompt-optimizer-v2

Pass

Audited by Gen Agent Trust Hub on Jul 22, 2026

Risk Level: SAFE
Full Analysis
  • [SAFE]: The skill provides comprehensive prompt engineering guidelines, including frameworks like RSCIT (Role, Situation, Constraints, Instructions, Template), chain-of-thought patterns, and few-shot examples. All instructions are pedagogical in nature and do not attempt to override agent safety filters or exfiltrate data.
  • [SAFE]: The metadata and provenance information in ORIGIN.md and metadata.json clearly identify the skill's source repository on GitHub. No obfuscation or malicious redirection patterns were found.
  • [SAFE]: The skill does not include any executable scripts, remote code downloads, or third-party package dependencies. It is a purely instruction-based skill.
  • [SAFE]: There are no attempts at privilege escalation or persistence mechanisms. The tool list in the frontmatter (codex-cli, claude-code, etc.) consists of standard developer tools and their inclusion here serves to scope the skill's environment without granting additional permissions.
Audit Metadata
Risk Level
SAFE
Analyzed
Jul 22, 2026, 04:05 PM
Security Audit — agent-trust-hub — llm-prompt-optimizer-v2