lov-i18n-check-i18n

Pass

Audited by Gen Agent Trust Hub on Sep 7, 2026

Risk Level: SAFEINDIRECT_PROMPT_INJECTIONCOMMAND_EXECUTION
Full Analysis
  • [INDIRECT_PROMPT_INJECTION]: The skill processes untrusted source code from the target project (JSX, placeholders, and UI text) to generate internationalization keys and update files. This provides an attack surface where malicious instructions embedded in the code could influence agent behavior.
  • Ingestion points: The workflow in SKILL.md explicitly scans user-visible text in JSX and notification files.
  • Boundary markers: SKILL.md includes an 'Execution boundary' section that mandates explicit request parsing and protects uncommitted content.
  • Capability inventory: The skill possesses capabilities for file system read/write operations and local command execution via provided utility scripts.
  • Sanitization: The instructions do not specify sanitization or validation of the scanned UI text before it is processed by the agent.
  • [COMMAND_EXECUTION]: The skill relies on local utility scripts, scripts/profile_store.py and scripts/validate_skill.py, to perform state management and source validation. These scripts are executed locally using the user's Python environment and interact with the ~/.config/agent-skills/ directory to persist non-sensitive user preferences and brand identity information.
Audit Metadata
Risk Level
SAFE
Analyzed
Sep 7, 2026, 09:47 AM
Security Audit — agent-trust-hub — lov-i18n-check-i18n