skills/crowdin/skills/crowdin-cli/Gen Agent Trust Hub

crowdin-cli

Pass

Audited by Gen Agent Trust Hub on Sep 21, 2026

Risk Level: SAFE
Full Analysis
  • [EXTERNAL_DOWNLOADS]: The skill provides instructions to install the Crowdin CLI from official sources such as NPM (@crowdin/cli), Homebrew, Docker, and GitHub releases. These are well-known and trusted vendor distribution channels.\n- [CREDENTIALS_UNSAFE]: The documentation guides the user on managing API tokens. It correctly highlights security best practices, such as using environment variables (CROWDIN_PERSONAL_TOKEN) and identity files (~/.crowdin.yml) instead of hardcoding secrets in version-controlled configuration files.\n- [INDIRECT_PROMPT_INJECTION]: The skill facilitates reading and writing localization strings from local project files. 1. Ingestion points: Local project files (globs in crowdin.yml). 2. Boundary markers: Absent in CLI output. 3. Capability inventory: upload, download, auto-translate, and string management. 4. Sanitization: Managed by the Crowdin platform and CLI logic. While this creates an ingestion surface for potentially untrusted data, it is the core functionality of the localization tool.
Audit Metadata
Risk Level
SAFE
Analyzed
Sep 21, 2026, 04:13 AM
Security Audit — agent-trust-hub — crowdin-cli