The Agent Skills Directory

[COMMAND_EXECUTION]: The skill utilizes standard development tools, specifically git diff to identify changes between documentation locales and pnpm --filter docs run build to verify the integrity of the generated documentation. These actions are restricted to the local repository environment.
[INDIRECT_PROMPT_INJECTION]: The skill ingests and processes content from documentation files located in docs/zh/, docs/en/, and docs/AGENTS.md. While processing external data is a potential injection surface, the skill defines strict terminology rules and structural constraints that limit the agent's interpretation of embedded instructions.
Ingestion points: Documentation files under docs/ and terminology rules in docs/AGENTS.md (SKILL.md).
Boundary markers: Not explicitly defined in the workflow.
Capability inventory: Git diff/syncing and pnpm build command execution (SKILL.md).
Sanitization: The skill explicitly instructs the agent to replace real identifiers with neutral placeholders during translation.
[CREDENTIALS_SAFE]: The instructions proactively mitigate data exposure by directing the agent to use placeholders such as example.com and YOUR_API_KEY instead of real internal values or credentials found in the source text.

translate-docs