skill-builder
Create Agent Skills with best practices, structure, validation, and testing. Use when designing or refining skills, prompts, references, or supporting files.
git clone --depth 1 https://github.com/existential-birds/beagle /tmp/skill-builder && cp -r /tmp/skill-builder/plugins/beagle-core/skills/skill-builder ~/.claude/skills/skill-builderSKILL.md
# Skill Builder Create, validate, and refine Agent Skills. ## Quick Start 1. Gather the capability, triggers, and required domain knowledge. 2. Choose a simple single-file skill or a multi-file skill with references. 3. Write `SKILL.md` with concise, trigger-focused instructions. 4. Add reference files only for detail that would otherwise bloat `SKILL.md`. 5. Validate YAML frontmatter, file layout, and naming. 6. Test the skill with the natural language users are likely to say. ## Workflow - Start with requirements and scope control. - Design the structure before writing content. - Keep descriptions in third person and include trigger keywords. - Use progressive disclosure for long examples, templates, and validation details. ## Gates Follow in order. **Pass** means a check you can satisfy with concrete artifacts (written bullets, paths, line counts, parsed YAML)—not an internal sense of “done.” 1. **Requirements** — **Pass:** Capability, triggers, and any required domain knowledge (or explicit “none”) are written down or confirmed from the user. 2. **Structure** — **Pass:** Single-file vs `SKILL.md` + `references/` is chosen; heavy detail lives in references, not inlined in `SKILL.md`. 3. **Draft** — **Pass:** Frontmatter is valid YAML with `name` and `description`; `SKILL.md` is ≤ 500 lines; every relative link from `SKILL.md` resolves to a path that exists under this skill directory; each `allowed-tools` entry (if present) is justified. 4. **Trigger check** — **Pass:** At least one natural-language user phrase plausibly matching the `description` is identified for a quick invocation test. ## Validation - Keep `SKILL.md` under 500 lines. - Prefer one-level reference links. - Avoid time-sensitive guidance. - Confirm frontmatter is valid YAML. - Check that any `allowed-tools` entries are necessary and correct. ## Advanced Reference For the full workflow, templates, examples, and validation checklist, see [references/skill-builder-guide.md](references/skill-builder-guide.md).
tag and push a release after the release PR is merged
create a release PR (auto-detects previous tag)
Guides architectural decisions for Deep Agents applications. Use when deciding between Deep Agents vs alternatives, choosing backend strategies, designing subagent systems, or selecting middleware approaches.
Reviews Deep Agents code for bugs, anti-patterns, and improvements. Use when reviewing code that uses create_deep_agent, backends, subagents, middleware, or human-in-the-loop patterns. Catches common configuration and usage mistakes.
Implements agents using Deep Agents. Use when building agents with create_deep_agent, configuring backends, defining subagents, adding middleware, or setting up human-in-the-loop workflows.
Guides architectural decisions for LangGraph applications. Use when deciding between LangGraph vs alternatives, choosing state management strategies, designing multi-agent systems, or selecting persistence and streaming approaches.
Reviews LangGraph code for bugs, anti-patterns, and improvements. Use when reviewing code that uses StateGraph, nodes, edges, checkpointing, or other LangGraph features. Catches common mistakes in state management, graph structure, and async patterns.
Implements stateful agent graphs using LangGraph. Use when building graphs, adding nodes/edges, defining state schemas, implementing checkpointing, handling interrupts, or creating multi-agent systems with LangGraph.