prompt-optimizer
Pass
Audited by Gen Agent Trust Hub on Oct 31, 2026
Risk Level: SAFENO_CODE
Full Analysis
- [SAFE]: The skill consists entirely of instructional text, diagrams, and methodologies for prompt optimization using the 4-D (Deconstruct, Diagnose, Develop, Deliver) framework.
- [NO_CODE]: There are no scripts, binaries, or executable commands included in the skill definition. The YAML frontmatter does not request any tool permissions or environment access.
- [INDIRECT_PROMPT_INJECTION]: The skill is designed to process user-supplied prompt text (referred to as 'Rough Input'). While this is an ingestion point for untrusted data, the skill's primary purpose is to apply structure (XML tags, delimiters) and clarity, which are recommended mitigations against instruction confusion and injection. The skill does not possess any capabilities (e.g., file system or network access) that could be exploited via indirect injection.
Audit Metadata