0langas-skill-center

Use after adding, renaming, importing, moving, or deleting a skill in 0langas-skill-center or any local agent customization repo. This skill reconciles canonical skill directories, provider discovery links, user-level Codex skill links, stale .agents copies, and local install state across Codex, Claude Code, and Kimi Code without duplicating content. Use proactively when skill visibility, symlinks, or local installs may drift.

Use when provider documentation has been consulted or should be reused for Codex, Claude Code, Kimi Code, plugin, skill, MCP, marketplace, config, hook, connector, install, or discovery work. This skill creates compact official-docs evidence notes so future agents do not repeat the same research or rely on memory. Use proactively after official docs lookup, before cross-provider claims, and when a task says docs were checked.

Use whenever working in 0langas-skill-center or another personal skill collection to add, rename, move, publish, document, test, or install skills. This skill orchestrates official-docs-citation-cache, tri-client-skill-port, skill-trigger-evalsmith, and agent-install-sync so every skill is canonical, documented, tri-client compatible, linked correctly, covered by activation evals, and safe to publish.

Use when a skill's autoactivation quality matters. This skill creates realistic should-trigger and should-not-trigger eval prompts, stores them beside the skill, checks whether descriptions are too broad or too weak, and helps tune trigger metadata without keyword spam. Use for new skills, renamed skills, cross-client ports, and skills that under-trigger or over-trigger in Codex, Claude Code, or Kimi Code.

Use when creating, importing, renaming, or repairing a skill that should work across Codex, Claude Code, and Kimi Code. This skill ports a canonical SKILL.md into a tri-client-compatible shape, verifies frontmatter fields, preserves provider-specific metadata, checks official docs evidence before discovery claims, and prepares provider discovery links without duplicating content.

Use this skill when the user wants a coding agent to work silently with near-zero chat during repository/codebase work. Trigger when the user writes "mute", "muteman", "/muteman", "no talking", "silent mode", "quiet mode", "no status updates", "shut up and code", "just work", "do not explain", or similar. The skill enforces a strict emoji-only communication protocol while preserving full internal planning, editing, testing, validation, recovery, and safety behavior.

Official documentation first for Codex, Claude Code, Kimi Code, and projects meant to work across multiple AI coding clients. Use automatically for provider configuration, config syntax, plugins, skills, connectors, MCP servers, hooks, subagents, memory/instruction files, auth setup, environment variables, CLI commands, IDE/desktop/web surfaces, model/provider routing, migrations, or compatibility between Codex, Claude Code, and Kimi Code. Check official docs before guessing provider-specific shapes or behavior.