doppler-convert

Pass

Audited by Gen Agent Trust Hub on Mar 12, 2026

Risk Level: SAFECOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
  • [COMMAND_EXECUTION]: The skill requires the agent to execute shell commands and local Node.js scripts for model conversion and debugging.\n
  • Evidence: Commands such as npm run convert, node tools/convert-safetensors-node.js, and npm run debug are specified in SKILL.md.\n
  • Context: These actions are necessary for the skill's primary function of model weight conversion and manifest verification.\n- [PROMPT_INJECTION]: There is a potential surface for indirect prompt injection due to the ingestion and review of external model data.\n
  • Ingestion points: External model files (Safetensors, GGUF) provided via the INPUT_PATH parameter in SKILL.md.\n
  • Boundary markers: The skill does not define specific delimiters or warnings to ignore embedded instructions within the ingested model data.\n
  • Capability inventory: The skill enables running inference via npm run debug and performing human-in-the-loop (HITL) review of the model's textual output.\n
  • Sanitization: No sanitization or escaping is performed on the output generated by the converted model before the agent evaluates its coherence.
Audit Metadata
Risk Level
SAFE
Analyzed
Mar 12, 2026, 10:48 PM