AGENT SKILL

setup-project

Use to generate HOTL adapter files for the current project — creates AGENTS.md, .clinerules, cursor rules, or copilot instructions depending on tools the team uses.

License 许可
Apache-2.0
最近更新
2026-11-14
Stars
★ 1.6k
Forks
491

01 前言 Front Matter

name
setup-project
description
Use to generate HOTL adapter files for the current project — creates AGENTS.md, .clinerules, cursor rules, or copilot instructions depending on tools the team uses.

03 SKILL.md 正文

Setup Project for HOTL

Overview

Generate the right config files so every code assistant on your team follows HOTL principles.

Announce: "Running HOTL project setup. Let me check what tools your team uses."

Process

  1. Ask: "Which code assistants does your team use?" (select all that apply)

    • Claude Code
    • Codex (OpenAI)
    • Cline (VS Code extension)
    • Cursor
    • GitHub Copilot
  2. Ask: "Will this project run multi-phase initiatives?" Default: no. Only answer yes for multi-phase work like major migrations, v1/v2 rewrites, platform rebuilds, or any effort that will span multiple phases with separate phase design docs and workflows. For a single feature, bug fix, or refactor, answer no — the standard HOTL flow handles those just fine.

    If yes, ask the follow-up: "What's the initiative slug? (kebab-case, e.g. ai-assurance, v2-migration)". Validate the answer matches [a-z0-9][a-z0-9-]*.

    If no (or if the user declines to opt in), skip step 5 entirely and proceed with tool-adapter generation only.

  3. For each selected tool, generate the appropriate file:

Tool File Generated Location
Claude Code CLAUDE.md Project root
Codex AGENTS.md; optional .codex/agents/*.toml Project root
Cline .clinerules Project root
Cursor .cursor/rules/hotl.md Project root
GitHub Copilot .github/copilot-instructions.md Project root
  1. Each generated file contains:

    • HOTL operating principles (intent/verification/governance contracts)
    • Brainstorming guidance that produces design docs in docs/designs/
    • Link to docs/plans/YYYY-MM-DD-<slug>-workflow.md format
    • Risk level guidelines
    • What always requires human review
  2. If Codex was selected, ask whether to install HOTL Codex custom agent templates. Default: yes. These templates make the bundled reviewer, architect, researcher, QA, dev, and PM roles available to the target project as Codex project-scoped custom agents.

    Resolve hotl-install-codex-agents.sh using the same six-location install-path order as the initiative scaffolder below:

    1. If you are working in the hotl-plugin repo itself, use scripts/hotl-install-codex-agents.sh
    2. Codex native-skills install: ~/.codex/hotl/scripts/hotl-install-codex-agents.sh
    3. Codex plugin install: ~/.codex/plugins/hotl-source/scripts/hotl-install-codex-agents.sh
    4. Codex plugin cache fallback: ~/.codex/plugins/cache/codex-plugins/hotl/*/scripts/hotl-install-codex-agents.sh
    5. Cline install fallback: ~/.cline/hotl/scripts/hotl-install-codex-agents.sh
    6. Claude Code plugin fallback: ~/.claude/plugins/hotl/scripts/hotl-install-codex-agents.sh

    Invoke the installer from the project root:

    bash <resolved-hotl-install-codex-agents.sh>
    

    The installer creates .codex/agents/ and copies the tracked HOTL templates from adapters/codex-agents/ into it. Existing agent files are preserved by default and emitted as SKIP: lines. Only use --force when the user explicitly approves overwriting existing project agent files.

  3. If the user opted in to initiative support in step 2, invoke the scaffolder to create .hotl/config.yml, the six docs/<tier>/ directories, and the four initiative-tier templates under docs/prompts/. The scaffolder should configure docs/designs/ as the canonical design-doc home and docs/plans/ as the canonical workflow home.

    Resolve hotl-init-initiative.sh using the same six-location order as document-lint.sh and hotl-config.sh (see skills/document-review/SKILL.md):

    1. If you are working in the hotl-plugin repo itself, use scripts/hotl-init-initiative.sh
    2. Codex native-skills install: ~/.codex/hotl/scripts/hotl-init-initiative.sh
    3. Codex plugin install: ~/.codex/plugins/hotl-source/scripts/hotl-init-initiative.sh
    4. Codex plugin cache fallback: ~/.codex/plugins/cache/codex-plugins/hotl/*/scripts/hotl-init-initiative.sh
    5. Cline install fallback: ~/.cline/hotl/scripts/hotl-init-initiative.sh
    6. Claude Code plugin fallback: ~/.claude/plugins/hotl/scripts/hotl-init-initiative.sh

    Invoke the scaffolder with the slug collected in step 2:

    bash <resolved-hotl-init-initiative.sh> --name <slug>
    

    The scaffolder refuses cleanly when .hotl/config.yml already exists — that is the intended behavior and should not be worked around. If any of the four target outputs under docs/prompts/ already exists, it is preserved byte-for-byte and a SKIP: line is emitted.

    Only invoke the scaffolder when the user answered yes in step 2. If the user answered no (the default), skip this step entirely.

  4. Commit all generated files:

git add AGENTS.md .clinerules .cursor/ .github/ CLAUDE.md
# If Codex custom agents were installed, also add:
#   .codex/agents/
# If initiative support was scaffolded in step 6, also add:
#   .hotl/config.yml docs/
git commit -m "chore: add HOTL adapter files for [tool list]"

AGENTS.md Template Content

# AGENTS.md — HOTL Operating Model

This project follows the Human-on-the-Loop (HOTL) development model.

## How to Work

1. Before feature work: brainstorm with intent/verification/governance contracts
2. Brainstorm into a design doc in `docs/designs/`, then create a dated workflow file at `docs/plans/YYYY-MM-DD-<slug>-workflow.md`
3. Execute steps autonomously within guardrails
4. Pause at `gate: human` for high-risk steps

## Risk Levels

- **low/medium + auto_approve: true:** Execute autonomously, auto-approve gates
- **high:** Always pause for human review at gates

## Always Requires Human Review

- Auth/authz changes
- Encryption or key management
- Privacy-critical logic (PII, consent, deletion)
- Billing or financial logic
- Broad access control changes

## Workflow Format

See `docs/plans/YYYY-MM-DD-<slug>-workflow.md` for canonical workflow instances or `workflows/` in the plugin for templates.

04 同仓库更多 Skills

hashgraph-online
hashgraph-online/awesome-codex-plugins

worktable-review-with-annotations

Create, inspect, reply to, update, or resolve situated Worktable annotations. Use when feedback belongs to a Doc section, block, text range, or HTML Doc, or when the user asks to address existing review feedback and coordinate a requested artifact edit.

★ 1.6k
hashgraph-online
hashgraph-online/awesome-codex-plugins

worktable-find-and-synthesize-context

Find, assess, and synthesize context from Worktable Docs and Records. Use when the user asks what a workspace contains, wants workspace-backed research, needs current facts reconciled across artifacts, or requests a cited synthesis without changing workspace state.

★ 1.6k
hashgraph-online
hashgraph-online/awesome-codex-plugins

worktable-create-or-update-docs

Create, edit, rename, or maintain narrative Worktable Docs. Use for plans, decisions, briefs, notes, research syntheses, and other durable prose when the user requests a new artifact or a safe update to an existing one.

★ 1.6k
hashgraph-online
hashgraph-online/awesome-codex-plugins

worktable-create-or-manage-records

Organize repeated structured information as durable Worktable Records. Use for research datasets, inventories, directories, pipelines, trackers, and other collections whose items need identity, fields, filtering, links, validation, or individual updates.

★ 1.6k
hashgraph-online
hashgraph-online/awesome-codex-plugins

worktable-create-or-manage-interactive-html

Create or improve durable visual and interactive artifacts in Worktable. Use for dashboards, trackers, calculators, mini apps, slide presentations, and other requests that benefit from a polished HTML Doc.

★ 1.6k
hashgraph-online
hashgraph-online/awesome-codex-plugins

worktable-collaborate-threads

Read, answer, and hand off messages through durable Worktable threads with users or other authenticated participants. Use when the user asks to check addressed messages, reply to a participant, continue a thread, request another agent's help, wait briefly for a response, or watch for comments and messages sent to the agent.

★ 1.6k