Skills
Navigate
skills/karakeep-app/karakeep/fill-translations/Gen Agent Trust Hub

fill-translations

Pass

Audited by Gen Agent Trust Hub on Oct 9, 2026

Risk Level: SAFEINDIRECT_PROMPT_INJECTIONCOMMAND_EXECUTION
Full Analysis
  • [INDIRECT_PROMPT_INJECTION]: The skill processes application translation strings which provide a surface for instructions to be embedded within data processed by the agent.\n
  • Ingestion points: Untrusted data is read from JSON files located at apps/web/lib/i18n/locales/*/translation.json via the i18n.mjs script.\n
  • Boundary markers: Absent. The agent is provided with keys and values to translate without delimiters or instructions to ignore embedded directives.\n
  • Capability inventory: The agent can modify the local filesystem (using the apply command) and execute git commands.\n
  • Sanitization: Absent. The strings collected for translation are passed to the agent without escaping or filtering.\n- [COMMAND_EXECUTION]: The skill utilizes a Node.js script that executes local system commands.\n
  • Evidence: The i18n.mjs script uses execFileSync and spawnSync from the node:child_process module to invoke git for repository root identification and version history analysis.
Audit Metadata
Risk Level
SAFE
Analyzed
Oct 9, 2026, 10:50 PM