prompt-optimizer

Pass

Audited by Gen Agent Trust Hub on Feb 17, 2026

Risk Level: SAFEPROMPT_INJECTIONNO_CODE
Full Analysis
  • [PROMPT_INJECTION] (LOW): The skill workflow facilitates indirect prompt injection by ingesting raw user prompts for optimization without establishing clear security boundaries.\n
  • Ingestion points: User input is received in SKILL.md (Step 1) for analysis and transformation.\n
  • Boundary markers: Absent. The skill does not employ delimiters (e.g., XML tags or triple quotes) to separate untrusted user content from the system instructions.\n
  • Capability inventory: The agent can generate and iteratively refine complex prompts, which could be exploited to bypass safety filters if malicious instructions are successfully injected.\n
  • Sanitization: The process relies on manual clarification (Step 4) but lacks specific instructions to identify or strip embedded commands from user-provided text.\n- [NO_CODE] (SAFE): The skill is composed entirely of Markdown documentation and reference files (59 files total). No executable scripts, binaries, or package manifests were found.
Audit Metadata
Risk Level
SAFE
Analyzed
Feb 17, 2026, 06:48 PM