token-saving-prompt-enhancer
Pass
Audited by Gen Agent Trust Hub on Sep 3, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONEXTERNAL_DOWNLOADS
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill acts as a prompt generator that ingests untrusted user data to produce instructions for an AI orchestrator. 1. Ingestion points: User-supplied rough ideas, requirements, or existing prompts processed in Phase 1 and 2. 2. Boundary markers: The generated prompt skeleton lacks explicit structural delimiters to separate user content from system instructions. 3. Capability inventory: The target environment (Claude Code) possesses significant capabilities including file system access and shell command execution. 4. Sanitization: No validation or sanitization of user-provided strings is mentioned before they are interpolated into the final prompt.
- [EXTERNAL_DOWNLOADS]: The skill's setup guide recommends the global installation of the @openai/codex package. This package is attributed to OpenAI, which is a recognized and well-known service, and its inclusion is documented as a functional component of the intended orchestration architecture.
Audit Metadata