article-translate
Warn
Audited by Gen Agent Trust Hub on Nov 12, 2026
Risk Level: MEDIUMDYNAMIC_EXECUTIONINDIRECT_PROMPT_INJECTIONCOMMAND_EXECUTION
Full Analysis
- [DYNAMIC_EXECUTION]: The script
scripts/translate.mjsdynamically imports and executes a Node.js module from a path constructed at runtime. The location of the shared engine is determined by either a relative path to a sibling directory (../../article2media/engine/scripts/) or a path provided via theARTICLE2MEDIA_ENGINEenvironment variable. This allows the skill to execute code from outside its own directory structure. - [INDIRECT_PROMPT_INJECTION]: The skill is designed to fetch and translate external article content (
article.md), which may contain malicious instructions intended to influence the agent's behavior during the translation process. - Ingestion points: The file
article.mdis processed, which can be populated by fetching content from arbitrary external URLs or files viaarticle-fetchorvideo-transcripttools. - Boundary markers: The skill requires block-by-block structural correspondence and uses a
checkscript to verify that the translated output matches the original structure, providing a limited integrity check against major structural deviations. - Capability inventory: The skill possesses the
Bashtool capability to execute Node.js scripts and has file system access to write translation files. - Sanitization: The instructions focus on structural consistency and terminology guidance, but there is no explicit filtering or sanitization of instructions embedded within the source article content.
- [COMMAND_EXECUTION]: The skill utilizes the
Bashtool to run Node.js utilities for generating article statistics and verifying the integrity of translated files. - Evidence: The
allowed-toolsconfiguration scopes execution tonode ${CLAUDE_SKILL_DIR}/scripts/*, and the skill body invokesnode ${CLAUDE_SKILL_DIR}/scripts/translate.mjsfor stats and check operations.
Audit Metadata