plugins/core/skills/manage-agent-instructions/SKILL.md
This skill should be used when creating, auditing, refactoring, or syncing CLAUDE.md and AGENTS.md files that configure AI coding agent behavior.
npx skillsauth add rbozydar/rbw-claude-code manage-agent-instructionsInstall this skill globally with one command. Works with Claude Code, Cursor, and Windsurf.
3 of 9 scanners reported clean
Some scanners were skipped, did not run, or reported a non-clean status. Review each row below.
<essential_principles>
These principles apply to ALL agent instruction files (CLAUDE.md, AGENTS.md).
The root instruction file should be as small as possible. Every token loads on every request regardless of relevance.
Absolute minimum content:
Everything else belongs in progressive disclosure files.
Structure knowledge in layers:
Level 1: Root file (~100-300 words)
Level 2: docs/CONVENTIONS.md, docs/TYPESCRIPT.md, etc.
Level 3: Nested references within those files
Load context only when needed for the current task.
File paths change constantly. Instead of "auth logic lives in src/auth/handlers.ts", write "Authentication is handled by the AuthService module" and let the agent discover current paths.
Curate ruthlessly. Remove redundant instructions (agent already knows), vague guidance ("write clean code"), contradicting rules, and auto-generated boilerplate.
Claude Code uses CLAUDE.md, not AGENTS.md. For multi-tool support, symlink or redirect between them.
| Level | Content | |-------|---------| | Root | Monorepo purpose, navigation, shared tools | | Package | Package purpose, specific stack, local conventions |
Keep each level focused on its scope. Avoid duplication.
</essential_principles>
<intake>What needs to be done?
Wait for response before proceeding.
</intake> <routing>| Response | Workflow |
|----------|----------|
| 1, "create", "new" | workflows/create-instructions.md |
| 2, "audit", "review" | workflows/audit-instructions.md |
| 3, "refactor", "shrink", "clean" | workflows/refactor-instructions.md |
| 4, "sync", "symlink" | workflows/sync-files.md |
| 5, "nested", "package", "monorepo" | workflows/nested-instructions.md |
After reading the workflow, follow it exactly.
</routing><templates_index>
Ready-to-use templates in templates/:
| Template | Use Case | |----------|----------| | minimal-claude-md.md | Standard project CLAUDE.md | | package-claude-md.md | Package-level instructions in a monorepo | | monorepo-root.md | Root-level monorepo instructions |
</templates_index>
<reference_index>
All in references/:
</reference_index>
<workflows_index>
All in workflows/:
| Workflow | Purpose | |----------|---------| | create-instructions.md | Create new CLAUDE.md or AGENTS.md from scratch | | audit-instructions.md | Analyze existing file for problems | | refactor-instructions.md | Fix bloated/problematic instruction files | | sync-files.md | Keep CLAUDE.md and AGENTS.md in sync | | nested-instructions.md | Add package-level instructions in monorepos |
</workflows_index>
<success_criteria>
A well-managed instruction file:
</success_criteria>
development
This skill should be loaded when writing, reviewing, or refactoring Python code to apply strict coding standards directly in the current context without spawning a subagent. It provides comprehensive Python development standards covering SOLID principles, asyncio patterns, type hints, testing, and production-quality code.
tools
This skill should be used when invoking the Gemini CLI for code review, plan review, or any prompt-based task. It provides correct invocation patterns emphasizing stdin piping and @ syntax over shell variable gymnastics.
development
Use this skill when writing, reviewing, or debugging Quickshell configurations (QML files for desktop shell UI on Wayland/Hyprland). Triggers on: QML files with Quickshell imports, shell.qml entry points, PanelWindow or FloatingWindow usage, Quickshell service integration (PipeWire, MPRIS, notifications, Hyprland IPC), Wayland layer-shell or session-lock code, custom bar/panel/widget/dock/OSD/lockscreen/launcher development, or any question about building a desktop shell with Quickshell on Hyprland.
development
This skill should be used when thorough, multi-perspective research with citations is needed. It performs comprehensive research using a diffusion research loop with domain specialization, supporting general web research and specialized domains (geopolitical with GDELT). Auto-detects domain from query or accepts an explicit --domain flag.