prompt-lab
Pass
Audited by Gen Agent Trust Hub on Sep 17, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill's workflow involves processing untrusted user prompts for analysis and optimization, which constitutes a surface for indirect prompt injection.
- Ingestion points: User-provided prompts are ingested in Phase 2 (Analyze Current Prompt) and Phase 3 (Generate Variants) as defined in SKILL.md.
- Boundary markers: The workflow lacks explicit instructions for the agent to use delimiters or boundary markers to isolate the analyzed prompt content from instructions.
- Capability inventory: The skill does not request any tools or network permissions in its frontmatter, and it primarily functions by generating text analysis and suggestions.
- Sanitization: No mechanisms for sanitizing or filtering the content of the user-provided prompts are described in the instructions.
Audit Metadata