french-tutor

Pass

Audited by Gen Agent Trust Hub on Apr 11, 2026

Risk Level: SAFECOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
  • [COMMAND_EXECUTION]: The skill facilitates integration with the Mochi SRS flashcard application by instructing the agent to execute a local Python script located at ~/.claude/skills/mochi-srs/scripts/mochi_api.py. This action is designed to be gated by user review and approval, minimizing the risk of unauthorized execution.
  • [PROMPT_INJECTION]: The skill processes untrusted user input in the form of French text for correction and analysis, which establishes an indirect prompt injection surface.
  • Ingestion points: User-provided text for correction in SKILL.md (Mode 3).
  • Boundary markers: Absent; the skill does not employ specific delimiters or instruction-guarding techniques for the ingested text.
  • Capability inventory: File system modification (state.json) and local command execution (mochi_api.py).
  • Sanitization: Absent; the skill relies on the agent's default safety guardrails to handle potentially malicious instructions within the French text.
  • [DATA_EXFILTRATION]: The skill maintains session state and vocabulary logs in a local directory (~/.claude/skills/french-tutor/state.json). This is standard state management behavior and does not involve any outbound network connections or sensitive system file access.
Audit Metadata
Risk Level
SAFE
Analyzed
Apr 11, 2026, 02:06 PM