apimart-generate-media

Fail

Audited by Gen Agent Trust Hub on Aug 12, 2026

Risk Level: CRITICALCOMMAND_EXECUTIONEXTERNAL_DOWNLOADS
Full Analysis
  • [COMMAND_EXECUTION]: The skill executes a bundled local Node.js script (scripts/apimart-media.mjs) to perform API operations such as listing models, fetching documentation, and submitting generation tasks.
  • [EXTERNAL_DOWNLOADS]: The skill communicates with external APIMart endpoints, including api.apimart.ai, api-dev.apimart.asia, and mcp.apimart.asia, to retrieve model specifications and documentation.
  • [DATA_EXFILTRATION]: The skill accesses the APIMART_API_KEY or API_KEY environment variables to authenticate requests. It implements strict guidelines to prevent these credentials from being exposed in chat, logs, or generated artifacts.
  • [INDIRECT_PROMPT_INJECTION]: The skill processes external Markdown documentation for AI models. It explicitly identifies this content as untrusted and instructs the agent to ignore any embedded instructions that attempt to reveal secrets or execute unrelated commands.
  • Ingestion points: External model documentation fetched via get_model_docs (MCP) or the docs command (local script).
  • Boundary markers: Explicit instructions in SKILL.md to treat documentation as untrusted and disregard embedded instructions.
  • Capability inventory: Network operations via fetch, local file reading via node:fs, and execution of the bundled utility script.
  • Sanitization: Logical filtering where the agent is instructed to only extract model parameters and ignore any action-oriented text within the reference materials.
Recommendations
  • CRITICAL: 1 file(s) identified as malware by FileRep - DO NOT USE
  • Contains 2 malicious URL(s) - DO NOT USE
Audit Metadata
Risk Level
CRITICAL
Analyzed
Aug 12, 2026, 01:58 AM