agentleFS
Sign inStart free
edwinidrusedwinidrus/dragonhoard/skills/init/SKILL.md

init

Skill2 starsChanged 9 days agoMITView on GitHub

Set up dragonhoard in the current project: ask the delivery mode (full Claude Code, or Claude Code orchestrating an OpenCode builder), ask where the implementation workspace lives, confirm the standard check commands, and scaffold the root files (constitution, status, protocol, logs, templates). Use when the user says "dragonhoard init", "set up dragonhoard", or invokes /dragonhoard:init. Run once per project.

137 other projects use a version of this.

Adopted by

Most recently taken up by

Used with

Projects with it also use daily-arxiv, exp-eval and 10 more.

Other files people use alongside this one.

Files found in the same public projects as a version of this one, more often than they turn up anywhere else. Each project counts once, large collections count for less, and this file's own repository is left out.
  • LambenthanSkill

    daily-arxiv

    每日从 arXiv 拉取新论文，过滤相关性，auto-ingest 高优先级论文，检测 SOTA 更新

    Also in 5 projects that use this

  • LambenthanSkill

    exp-eval

    实验判决门：Review LLM 独立评判实验结果 → 4 种判决路径 → 自动更新 claims confidence、ideas status、graph edges

    Also in 5 projects that use this

  • skyllwtSkill

    exp-pilot-eval

    Pilot result evaluation — read pilot results, apply success criteria, update idea page (pilot_result, failure_reason if failed), generate PILOT_VERDICT_REPORT. Called by /ideate Phase 5 after /exp-pilot-run.

    Also in 4 projects that use this

  • skyllwtSkill

    exp-pilot-run

    Pilot experiment execution — read Pilot Spec YAML, write pilot code, run experiment(Confirm with the user before operation and require the applicant to conduct manual inspection), return results. Called by /ideate Phase 5. Does NOT modify wiki pages or judge pass/fail.

    Also in 4 projects that use this

  • gaebalaiSkill

    codex-system

    PROACTIVELY consult Codex CLI, your highly capable supporter with exceptional reasoning and task completion abilities. Codex is a trusted expert you should ALWAYS consult BEFORE making decisions on: design choices, implementation approaches, debugging strategies, refactoring plans, or any non-trivial problem. When uncertain, consult Codex. Don't hesitate - Codex provides better analysis. Explicit triggers: "think deeper", "analyze", "second opinion", "consult codex".

    Also in 3 projects that use this

  • +6 more from Lambenthan
  • +1 more from skyllwt

The file

We split this skill into the rules it gives an agent

6 rules. 0 are common, 1 is distinctive, and 5 are very specific to this project.

  • 0

    Common

    Most of the 125 edited versions we read keep these.

  • 1

    Distinctive

    This author's own choices. Worth a look.

  • 5

    Very specific

    Tied to their project's names, files or tools. Probably skip.

  • Common
  • Distinctive
  • Very specific
  1. Very specific:

    Already set up?: If dragonhoard.json at the project root already has a root, report it and stop.

    line 15
  2. Very specific:

    Ask the delivery mode, always: Ask the owner and wait (in Claude Code, use AskUserQuestion).

    line 19
  3. Distinctive:

    Ask the root folder, always: Do not assume one. Ask the owner and wait for the answer (in Claude Code, use AskUserQuestion).

    line 30
  4. Very specific:

    Detect check commands: Read only the manifests that exist: pyproject.toml, package.json scripts, Makefile, Cargo.toml, go.mod, CI files …

    line 36
  5. Very specific:

    Project articles: Ask whether the project has hard constraints worth adding to <root>/CONSTITUTION.md.

    line 52
  6. Very specific:

    Role models: Run /dragonhoard:models. The current model suggests a model for each role, and the owner confirms.

    line 71

“Common” is measured: kept by at least 6 in 10 of the 125 edited versions of this file we could read. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.

Keep what fits and drop the rest. Your version, ready for your agents.
Files
SKILL.md

dragonhoard init

CLI: python3 "${CLAUDE_PLUGIN_ROOT}/scripts/dragonhoard.py" (called dragonhoard.py below).

  1. Already set up? If dragonhoard.json at the project root already has a root, report it and stop. Never overwrite it. A dragonhoard.json holding only host (and models) was written by install; continue.

  2. Ask the delivery mode, always. Ask the owner and wait (in Claude Code, use AskUserQuestion). Offer exactly two options, with the trade-off from docs/modes.md:

    • Full Claude Code (solo, recommended when quality matters most and budget is fine). This session builds; scout, spiker and judge run as Claude subagents.
    • Claude Code + OpenCode (hybrid, recommended to save Claude tokens). This session orchestrates, judges and verifies. A headless OpenCode session builds each task on a cheaper or free model. Requires opencode on PATH: check opencode --version before offering it, and say so if it is missing.

    Outside Claude Code there is nothing to orchestrate from, so use solo without asking.

  3. Ask the root folder, always. Do not assume one. Ask the owner and wait for the answer (in Claude Code, use AskUserQuestion). Offer:
    • implementation/ — building a plan from scratch.
    • improvement/ — improving an existing system.
    • Any other relative folder the owner names.
  4. Detect check commands. Read only the manifests that exist: pyproject.toml, package.json scripts, Makefile, Cargo.toml, go.mod, CI files under .github/workflows/. Propose the lint, format-check, type-check and test commands the project already uses. Never invent a tool the project does not have. Ask the owner to confirm or edit the proposed set, and wait. Zero checks is allowed for a project with no code yet.

  5. Scaffold.

    python3 "${CLAUDE_PLUGIN_ROOT}/scripts/dragonhoard.py" init --root <root> --mode <solo|hybrid> --check "<cmd 1>" --check "<cmd 2>"

    It writes dragonhoard.json, then <root>/README.md, CONSTITUTION.md, STATUS.md, LESSONS.md and DIGEST.md, then <root>/logs/ and <root>/templates/. It never overwrites an existing file.

  6. Project articles. Ask whether the project has hard constraints worth adding to <root>/CONSTITUTION.md. Examples: allowed dependencies, data conventions, a storage choice, security rules. Append each one the owner confirms as a new numbered article after section VII. Do not add generic advice.
  7. Hybrid only: set up the builder.
    • List the MCP servers OpenCode would load (opencode mcp list). Ask the owner which ones the builder may use. The default is none that reach personal data such as notes, mail or drives.

    • Install, passing the rest as --mcp-off:

      dragonhoard.py install --host opencode --mcp-off <name,name>

      This writes .dragonhoard/, the dragonhoard-* skills, the role agents, AGENTS.md, and opencode.json. In opencode.json git commits are denied, no rule ever asks (a headless run would hang), and only the dragonhoard and discipline skills are allowed.

    • The judge stays external: you are the judge.
  8. Role models. Run /dragonhoard:models. The current model suggests a model for each role, and the owner confirms. Solo has scout, spiker and judge. Hybrid has builder, scout and spiker; for those, see opencode models, where models marked free cost nothing.
  9. Hand off. Show the dragon banner the command printed. Then suggest /dragonhoard:shape <plan file> as the next step. In hybrid, shaping also runs through /dragonhoard:conduct, so the builder writes the packs and you review them.

Download now (ZIP)

Does it work?

Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.

No reviews yet

People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.

  • It worked
  • Partly
  • It didn't
What did you use it for? e.g. Reviewing pull requests in a Rails app

Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.

Your trail

Files you open will collect here, so you can find them again.

Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.

Now do this for your own.

Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.

Check your setup, free