prompt-improver
Pass
Audited by Gen Agent Trust Hub on Mar 9, 2026
Risk Level: SAFENO_CODEPROMPT_INJECTION
Full Analysis
- [NO_CODE]: The skill package consists strictly of Markdown documentation and a manifest. No Python, JavaScript, or shell scripts are included, and no automated tools are defined.
- [PROMPT_INJECTION]: The skill's primary function involves analyzing untrusted user input (prompts to be improved). While this represents a surface for indirect prompt injection, the skill lacks any functional capabilities such as network access, file system modifications, or tool execution to act upon malicious instructions. Ingestion points: user-provided prompt text. Boundary markers: absent. Capability inventory: text generation only. Sanitization: none.
Audit Metadata