The Agent Skills Directory

[EXTERNAL_DOWNLOADS]: The scripts/sync-language-registry.mjs file fetches linguistic data from the Indonesian Ministry of Education's official portal (petabahasa.kemendikdasmen.go.id) to synchronize the local language registry.
[COMMAND_EXECUTION]: Development and test scripts utilize child_process.spawnSync to execute local Node.js utilities for output validation and skill linting during the development phase.
[INDIRECT_PROMPT_INJECTION]: The skill processes user input to apply regional voices and poetic styles. It mitigates potential injection risks by explicitly instructing the agent in references/core.md to preserve technical artifacts (code, commands, paths) and to never allow style changes to override security warnings or permission decisions.
Ingestion points: evals/cases.json and user-supplied prompts.
Boundary markers: Explicit instructions in references/core.md regarding 'artefak terlindungi' (protected artifacts) to maintain technical integrity.
Capability inventory: Local file system access and subprocess spawning for testing purposes in scripts/.
Sanitization: Validation scripts check for the preservation of key technical tokens and the absence of restricted linguistic patterns.

bahasa-indonesia