notemdpro-text-translator

Pass

Audited by Gen Agent Trust Hub on Apr 16, 2026

Risk Level: SAFEPROMPT_INJECTION
Full Analysis
  • [PROMPT_INJECTION]: The skill exhibits a surface for indirect prompt injection.
  • Ingestion points: Content from local markdown files is read into the agent context via the read_file function described in SKILL.md.
  • Boundary markers: The prompt template in promptUtils.ts uses a simple 'Text to translate:' label but lacks structural delimiters or explicit instructions to disregard any commands embedded in the note content.
  • Capability inventory: The skill has filesystem access to read and write files (read_file, write_file).
  • Sanitization: No sanitization or escaping of the input {TEXT} variable is documented or implemented before it is interpolated into the prompt.
Audit Metadata
Risk Level
SAFE
Analyzed
Apr 16, 2026, 02:49 PM