dby-charter

Pass

Audited by Gen Agent Trust Hub on Sep 9, 2026

Risk Level: SAFEINDIRECT_PROMPT_INJECTIONCOMMAND_EXECUTION
Full Analysis
  • [INDIRECT_PROMPT_INJECTION]: The style distillation feature processes untrusted text samples, which can be used to influence agent behavior.
  • Ingestion points: references/writing-dna.md allows users to submit writing samples for analysis.
  • Boundary markers: Untrusted samples are wrapped in <<<SAMPLE n>>> and <<<END SAMPLE n>>> tags.
  • Capability inventory: The skill includes scripts/charter.mjs for file access and network requests to doubaoya.com, and uses the DOUBAOYA_API_KEY for authentication.
  • Sanitization: The prompt explicitly instructs the LLM to ignore any instructions found within the sample data.
  • [COMMAND_EXECUTION]: The skill uses a local Node.js script, scripts/charter.mjs, to handle API requests and local file operations.
  • The script reads local JSON files for updates and communicates with the vendor's official domain at https://doubaoya.com.
  • Secret handling is managed through environment variables rather than hardcoding.
Audit Metadata
Risk Level
SAFE
Analyzed
Sep 9, 2026, 02:51 PM
Security Audit — agent-trust-hub — dby-charter