{"id":40948998,"url":"https://github.com/minipuft/claude-prompts","last_synced_at":"2026-04-01T18:06:15.538Z","repository":{"id":280334890,"uuid":"941652610","full_name":"minipuft/claude-prompts","owner":"minipuft","description":"MCP prompt template server: hot-reload, thinking frameworks, quality gates","archived":false,"fork":false,"pushed_at":"2026-03-25T09:29:13.000Z","size":40892,"stargazers_count":143,"open_issues_count":13,"forks_count":30,"subscribers_count":3,"default_branch":"main","last_synced_at":"2026-03-25T09:35:11.264Z","etag":null,"topics":["ai-automation","ai-development","ai-workflows","claude","llm","llm-tools","mcp","model-context-protocol","nodejs","prompt-engineering","prompt-management","prompt-templates","typescript"],"latest_commit_sha":null,"homepage":"","language":"TypeScript","has_issues":true,"has_wiki":null,"has_pages":null,"mirror_url":null,"source_name":null,"license":"agpl-3.0","status":null,"scm":"git","pull_requests_enabled":true,"icon_url":"https://github.com/minipuft.png","metadata":{"files":{"readme":"README.md","changelog":"CHANGELOG.md","contributing":"CONTRIBUTING.md","funding":null,"license":"LICENSE","code_of_conduct":null,"threat_model":null,"audit":null,"citation":null,"codeowners":null,"security":null,"support":null,"governance":null,"roadmap":null,"authors":null,"dei":null,"publiccode":null,"codemeta":null,"zenodo":null,"notice":null,"maintainers":null,"copyright":null,"agents":null,"dco":null,"cla":null}},"created_at":"2025-03-02T19:36:08.000Z","updated_at":"2026-03-24T07:51:43.000Z","dependencies_parsed_at":"2025-07-10T16:11:43.492Z","dependency_job_id":"bc5b45bf-0017-4687-b1c8-a33e9c33bb42","html_url":"https://github.com/minipuft/claude-prompts","commit_stats":null,"previous_names":["minipuft/claude-prompts","minipuft/claude-prompts-mcp"],"tags_count":15,"template":false,"template_full_name":null,"purl":"pkg:github/minipuft/claude-prompts","repository_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/minipuft%2Fclaude-prompts","tags_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/minipuft%2Fclaude-prompts/tags","releases_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/minipuft%2Fclaude-prompts/releases","manifests_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/minipuft%2Fclaude-prompts/manifests","owner_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/owners/minipuft","download_url":"https://codeload.github.com/minipuft/claude-prompts/tar.gz/refs/heads/main","sbom_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/minipuft%2Fclaude-prompts/sbom","scorecard":null,"host":{"name":"GitHub","url":"https://github.com","kind":"github","repositories_count":286080680,"owners_count":31290749,"icon_url":"https://github.com/github.png","version":null,"created_at":"2022-05-30T11:31:42.601Z","updated_at":"2026-04-01T13:12:26.723Z","status":"ssl_error","status_checked_at":"2026-04-01T13:12:25.102Z","response_time":53,"last_error":"SSL_read: unexpected eof while reading","robots_txt_status":"success","robots_txt_updated_at":"2025-07-24T06:49:26.215Z","robots_txt_url":"https://github.com/robots.txt","online":false,"can_crawl_api":true,"host_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub","repositories_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories","repository_names_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repository_names","owners_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/owners"}},"keywords":["ai-automation","ai-development","ai-workflows","claude","llm","llm-tools","mcp","model-context-protocol","nodejs","prompt-engineering","prompt-management","prompt-templates","typescript"],"created_at":"2026-01-22T05:08:47.530Z","updated_at":"2026-04-01T18:06:15.525Z","avatar_url":"https://github.com/minipuft.png","language":"TypeScript","funding_links":[],"categories":["TypeScript"],"sub_categories":[],"readme":"# Claude Prompts MCP Server\n\n\u003cdiv align=\"center\"\u003e\n\n\u003cimg src=\"assets/logo.png\" alt=\"Claude Prompts MCP Server Logo\" width=\"200\" /\u003e\n\n[![npm version](https://img.shields.io/npm/v/claude-prompts.svg?style=for-the-badge\u0026logo=npm\u0026color=0066cc)](https://www.npmjs.com/package/claude-prompts)\n[![License: AGPL v3](https://img.shields.io/badge/License-AGPL_v3-00ff88.svg?style=for-the-badge\u0026logo=gnu)](https://www.gnu.org/licenses/agpl-3.0)\n\n**An MCP workflow server.**\n\nCraft reusable prompts with validation and reasoning guidance.\u003cbr\u003e\nOrchestrate agentic workflows with a composable operator syntax.\u003cbr\u003e\nExport as native skills.\n\n[Quick Start](#quick-start) · [What You Get](#what-you-get) · [Compose Workflows](#compose-workflows) · [Run Anywhere](#run-anywhere) · [Docs](#documentation)\n\n\u003c/div\u003e\n\n\u003cdiv align=\"center\"\u003e\n\n\u003cimg src=\"assets/demos/hero-demo.gif\" alt=\"Chain workflow with gate validation — prompt executes through hooks, gate catches missing field on first attempt then self-corrects\" width=\"720\" /\u003e\n\n\u003csub\u003eChain + gate validation in action (haiku model) — gates catch errors and guide self-correction, even on the cheapest model\u003c/sub\u003e\n\n\u003c/div\u003e\n\n### What your AI client gives you — and what this server adds\n\n| Your client already does   | This server adds                                                                      |\n| -------------------------- | ------------------------------------------------------------------------------------- |\n| Run a prompt               | Compose prompts with validation, reasoning guidance, and formatting in one expression |\n| Single-shot skills         | Multi-step workflows that thread context between steps                                |\n| Execute subagents          | Hand off mid-chain steps to agents with full workflow context                         |\n| Client-native skill format | Author once as YAML, export to any client with `skills:export`                        |\n| Manual prompt writing      | Versioned templates with hot-reload, rollback, and history                            |\n| Trust the output           | Validate output between steps — self-evaluation and shell commands                    |\n\n---\n\n## Quick Start\n\n### Claude Code (Recommended)\n\n```bash\n# Add marketplace (first time only)\n/plugin marketplace add minipuft/minipuft-plugins\n\n# Install\n/plugin install claude-prompts@minipuft\n\n# Try it\n\u003e\u003etech_evaluation_chain library:'zod' context:'API validation'\n```\n\n\u003cdetails\u003e\n\u003csummary\u003eDevelopment setup\u003c/summary\u003e\n\nLoad plugin from local source for development:\n\n```bash\ngit clone https://github.com/minipuft/claude-prompts ~/Applications/claude-prompts\ncd ~/Applications/claude-prompts/server \u0026\u0026 npm install \u0026\u0026 npm run build\nclaude --plugin-dir ~/Applications/claude-prompts\n```\n\nEdit hooks/prompts → restart Claude Code. Edit TypeScript → rebuild first.\n\n\u003c/details\u003e\n\n**Custom prompts**: Use `--init=~/my-prompts` to create a workspace with starter templates you own. Prompts created via `resource_manager` are saved to your active resources directory. See [Custom Resources](#custom-resources).\n\n---\n\n### Claude Desktop\n\n**Option A: GitHub Release** (recommended)\n\n1. Download `claude-prompts-{version}.mcpb` from [Releases](https://github.com/minipuft/claude-prompts/releases/latest)\n2. Drag into Claude Desktop Settings → MCP Servers\n3. Done\n\nThe `.mcpb` bundle is self-contained (~5MB) — no npm required.\n\n**Option B: NPX** (auto-updates)\n\nAdd to your config file:\n\n- macOS: `~/Library/Application Support/Claude/claude_desktop_config.json`\n- Windows: `%APPDATA%\\Claude\\claude_desktop_config.json`\n\n```json\n{\n  \"mcpServers\": {\n    \"claude-prompts\": {\n      \"command\": \"npx\",\n      \"args\": [\"-y\", \"claude-prompts@latest\", \"--client\", \"claude-code\"]\n    }\n  }\n}\n```\n\nRestart Claude Desktop and test: `\u003e\u003eresearch_chain topic:'remote team policies'`\n\n---\n\n\u003cdetails\u003e\n\u003csummary\u003e\u003cstrong\u003eVS Code / Copilot\u003c/strong\u003e\u003c/summary\u003e\n\n[![Install in VS Code](https://img.shields.io/badge/VS_Code-Install_Server-0098FF?style=flat-square\u0026logo=visualstudiocode\u0026logoColor=white)](https://insiders.vscode.dev/redirect/mcp/install?name=claude-prompts\u0026config=%7B%22command%22%3A%22npx%22%2C%22args%22%3A%5B%22-y%22%2C%22claude-prompts%40latest%22%5D%7D\u0026quality=stable)\n\nClick the badge above for one-click install, or add manually to `.vscode/mcp.json`:\n\n```json\n{\n  \"servers\": {\n    \"claude-prompts\": {\n      \"command\": \"npx\",\n      \"args\": [\"-y\", \"claude-prompts@latest\"]\n    }\n  }\n}\n```\n\n\u003c/details\u003e\n\n\u003cdetails\u003e\n\u003csummary\u003e\u003cstrong\u003eCursor\u003c/strong\u003e\u003c/summary\u003e\n\n[![Install in Cursor](https://img.shields.io/badge/Cursor-Install_Server-F14F21?style=flat-square\u0026logo=cursor\u0026logoColor=white)](cursor://anysphere.cursor-deeplink/mcp/install?name=claude-prompts\u0026config=eyJjb21tYW5kIjoibnB4IiwiYXJncyI6WyIteSIsImNsYXVkZS1wcm9tcHRzQGxhdGVzdCJdfQ==)\n\nClick the badge above for one-click install, or add manually to `~/.cursor/mcp.json`:\n\n```json\n{\n  \"mcpServers\": {\n    \"claude-prompts\": {\n      \"command\": \"npx\",\n      \"args\": [\"-y\", \"claude-prompts@latest\", \"--client=cursor\"]\n    }\n  }\n}\n```\n\n\u003c/details\u003e\n\n\u003cdetails\u003e\n\u003csummary\u003e\u003cstrong\u003eOpenCode\u003c/strong\u003e\u003c/summary\u003e\n\n**Option A: Plugin install** (recommended — includes hooks)\n\n```bash\nnpm install -g opencode-prompts\nopencode-prompts install\n```\n\nThe installer configures hooks (chain tracking, gate enforcement, state preservation), plugin registration, and MCP server. See [opencode-prompts](https://github.com/minipuft/opencode-prompts) for what hooks provide.\n\n**Option B: MCP server only** (manual config, no hooks)\n\nAdd to `~/.config/opencode/opencode.json`:\n\n```json\n{\n  \"mcp\": {\n    \"claude-prompts\": {\n      \"type\": \"local\",\n      \"command\": [\n        \"npx\",\n        \"-y\",\n        \"claude-prompts@latest\",\n        \"--transport=stdio\",\n        \"--client=opencode\"\n      ]\n    }\n  }\n}\n```\n\nYou'll have the three MCP tools but no chain tracking, gate enforcement, or state preservation across compactions. To load your own prompts, add an `environment` key — see [Custom Resources](#custom-resources).\n\n\u003c/details\u003e\n\n\u003cdetails\u003e\n\u003csummary\u003e\u003cstrong\u003eGemini CLI\u003c/strong\u003e\u003c/summary\u003e\n\n**Option A: Extension install** (recommended — includes hooks)\n\n```bash\n# Enable hooks (first time only — skip if already enabled)\necho '{\"hooks\": {\"enabled\": true}}' \u003e ~/.gemini/settings.json\n\n# Install extension\ngemini extensions install https://github.com/minipuft/gemini-prompts\n```\n\nThe extension registers the MCP server and adds hooks for `\u003e\u003e` syntax detection, chain tracking, and gate reminders. See [gemini-prompts](https://github.com/minipuft/gemini-prompts) for what hooks provide.\n\n**Option B: MCP server only** (manual config, no hooks)\n\nAdd to `~/.gemini/settings.json`:\n\n```json\n{\n  \"mcpServers\": {\n    \"claude-prompts\": {\n      \"command\": \"npx\",\n      \"args\": [\"-y\", \"claude-prompts@latest\", \"--client=gemini\"]\n    }\n  }\n}\n```\n\nYou'll have the three MCP tools but no `\u003e\u003e` syntax detection, chain tracking, or gate reminders. To load your own prompts, add an `env` key — see [Custom Resources](#custom-resources).\n\n\u003c/details\u003e\n\n\u003cdetails\u003e\n\u003csummary\u003e\u003cstrong\u003eOther Clients\u003c/strong\u003e (Codex, Windsurf, Zed)\u003c/summary\u003e\n\nAdd to your MCP config file with a `--client` preset for deterministic handoff guidance:\n\n| Client   | Config Location                           | Recommended `--client`  |\n| -------- | ----------------------------------------- | ----------------------- |\n| Codex    | `~/.codex/config.toml`                    | `codex`                 |\n| Windsurf | `~/.codeium/windsurf/mcp_config.json`     | `cursor` (experimental) |\n| Zed      | `~/.config/zed/settings.json` → `mcp` key | `unknown`               |\n\n**JSON-based configs (Windsurf/Zed):**\n\n```json\n{\n  \"mcpServers\": {\n    \"claude-prompts\": {\n      \"command\": \"npx\",\n      \"args\": [\"-y\", \"claude-prompts@latest\", \"--client=cursor\"]\n    }\n  }\n}\n```\n\n**Codex (`~/.codex/config.toml`):**\n\n```toml\n[mcp_servers.claude_prompts]\ncommand = \"npx\"\nargs = [\"-y\", \"claude-prompts@latest\", \"--client=codex\"]\n```\n\nSupported presets: `claude-code`, `codex`, `gemini`, `opencode`, `cursor`, `unknown`.\n\nFor complete per-client setup and limitations:\n\n- [Client Integration Guide](docs/guides/client-integration.md)\n- [Client Capabilities Reference](docs/reference/client-capabilities.md)\n\n\u003c/details\u003e\n\n\u003cdetails\u003e\n\u003csummary\u003e\u003cstrong\u003eFrom Source\u003c/strong\u003e (developers only)\u003c/summary\u003e\n\n```bash\ngit clone https://github.com/minipuft/claude-prompts.git\ncd claude-prompts/server\nnpm install \u0026\u0026 npm run build \u0026\u0026 npm test\n```\n\nPoint your MCP config to `server/dist/index.js`. The esbuild bundle is self-contained.\n\n**Transport options**: `--transport=stdio` (default), `--transport=streamable-http` (HTTP clients).\n\n\u003c/details\u003e\n\n### Custom Resources\n\n**Use your own prompts, gates, methodologies, and styles.** Two approaches depending on whether you want the bundled resources or not:\n\n**Option A: Own workspace** (recommended for full control)\n\nCreate a workspace with starter templates, then point your MCP config to it:\n\n```bash\nnpx -y claude-prompts@latest --init=~/my-prompts\n```\n\nThis creates `~/my-prompts/resources/` with starter prompts you own. Set `MCP_WORKSPACE` or `MCP_RESOURCES_PATH` to use it. Prompts created via `resource_manager` are saved here. Your AI can update them through MCP — no manual editing needed.\n\n**Option B: Plugin install** (bundled resources + hooks)\n\nPlugin installs (Claude Code, OpenCode, Gemini) set `MCP_WORKSPACE` automatically and ship the bundled 90+ prompts, gates, and methodologies. Prompts created via `resource_manager` are saved to the plugin's resources directory.\n\n\u003e [!IMPORTANT]\n\u003e **`MCP_RESOURCES_PATH`** sets the base resources directory (replaces the package default).\n\u003e **`MCP_WORKSPACE`** enables overlay — custom resources in your workspace are loaded **alongside** bundled ones. Prompts, gates, methodologies, and styles with the same ID as bundled ones take priority.\n\n\u003cdetails\u003e\n\u003csummary\u003e\u003cstrong\u003eConfig examples per client\u003c/strong\u003e\u003c/summary\u003e\n\n**Claude Desktop / VS Code / Cursor** (JSON with `env`):\n\n```json\n{\n  \"mcpServers\": {\n    \"claude-prompts\": {\n      \"command\": \"npx\",\n      \"args\": [\"-y\", \"claude-prompts@latest\"],\n      \"env\": {\n        \"MCP_RESOURCES_PATH\": \"/path/to/your/resources\"\n      }\n    }\n  }\n}\n```\n\n**OpenCode** (JSON with `environment`):\n\n```json\n{\n  \"mcp\": {\n    \"claude-prompts\": {\n      \"type\": \"local\",\n      \"command\": [\"npx\", \"-y\", \"claude-prompts@latest\", \"--transport=stdio\"],\n      \"environment\": {\n        \"MCP_RESOURCES_PATH\": \"/path/to/your/resources\"\n      }\n    }\n  }\n}\n```\n\n\u003c/details\u003e\n\nSee [CLI Configuration](docs/reference/mcp-tools.md#cli-configuration) for all env vars including fine-grained path overrides (`MCP_PROMPTS_PATH`, `MCP_GATES_PATH`, etc.).\n\n---\n\n\u003cdetails\u003e\n\u003csummary\u003e\u003cstrong\u003eSee the dashboard\u003c/strong\u003e — system status overview\u003c/summary\u003e\n\n\u003cbr\u003e\n\n\u003cimg src=\"assets/demos/status-demo-3x.gif\" alt=\"System status demo showing loaded prompts, gates, methodologies, and active configuration\" width=\"720\" /\u003e\n\n\u003csub\u003eLoaded resources, active configuration, and server health at a glance\u003c/sub\u003e\n\n\u003c/details\u003e\n\n---\n\n## What You Get\n\nFour resource types you author, version, and compose into workflows.\n\n\u003cdetails\u003e\n\u003csummary\u003e\u003cstrong\u003eSee the catalog\u003c/strong\u003e — listing all available prompts\u003c/summary\u003e\n\n\u003cbr\u003e\n\n\u003cimg src=\"assets/demos/resource-list-demo.gif\" alt=\"Listing all available prompts across 11 categories using the resource_manager tool\" width=\"720\" /\u003e\n\n\u003csub\u003e90 prompts across 11 categories — all hot-reloadable and versionable\u003c/sub\u003e\n\n\u003c/details\u003e\n\n### Prompt Templates\n\nVersioned YAML with hot-reload. Edit a template, test it immediately — or ask your AI to update it through MCP.\n\n```\n\u003e\u003ecode_review target:'src/auth/' language:'typescript'\n```\n\n### Validation Rules (Gates)\n\nCriteria the AI checks its own output against. Blocking or advisory.\n\n```\n:: 'no false positives' :: 'cite sources with links'\n```\n\nFailed checks can retry automatically or pause for your decision.\n\n\u003e [!TIP]\n\u003e **Define your own checks.** See the [Gates Guide](docs/guides/gates.md) for blocking vs advisory rules, retry behavior, and shell verification.\n\n### Reasoning Guidance (Methodologies)\n\nFrameworks that shape how the AI thinks through a problem — not just what it outputs. 6 built-in, or create your own.\n\n```\n@CAGEERF    # Context → Analysis → Goals → Execution → Evaluation → Refinement\n@ReACT      # Reason → Act → Observe loops\n@5W1H       # Who, What, Where, When, Why, How\n```\n\n\u003e [!TIP]\n\u003e **Create your own framework.** See the [Methodologies Guide](docs/guides/methodologies.md) for built-in frameworks and custom authoring.\n\n### Styles\n\nResponse formatting and tone.\n\n```\n#analytical    # Structured, evidence-based output\n#concise       # Brief, action-focused\n```\n\nAll resources are hot-reloadable, versioned with rollback history, and managed through the `resource_manager` tool.\n\n\u003e [!TIP]\n\u003e **Ready to build your own?** Start with the [Prompt Authoring Tutorial](docs/tutorials/build-first-prompt.md).\n\n---\n\n## Compose Workflows\n\nThe operator syntax wires resources together — chain steps, add validation inline, hand off steps to agents.\n\n```\n\u003e\u003ereview target:'src/auth/' @CAGEERF :: 'no false positives'\n  --\u003e security_scan :: verify:\"npm test\"\n  --\u003e recommendations :: 'actionable, with code'\n  ==\u003e implementation\n```\n\n\u003cdetails\u003e\n\u003csummary\u003e\u003cstrong\u003eSee the chain\u003c/strong\u003e — phases completing back-to-back\u003c/summary\u003e\n\n\u003cbr\u003e\n\n\u003cimg src=\"assets/demos/hero-chain-demo.gif\" alt=\"Chain phases 3-4 executing back-to-back, compounding reasoning across steps before rendering final output\" width=\"720\" /\u003e\n\n\u003csub\u003ePhases compound reasoning across steps — each step builds on validated output from the previous one\u003c/sub\u003e\n\n\u003c/details\u003e\n\n\u003cdetails\u003e\n\u003csummary\u003e\u003cstrong\u003eSee the output\u003c/strong\u003e — tech evaluation chain with context7 research\u003c/summary\u003e\n\n\u003cbr\u003e\n\n\u003cimg src=\"assets/demos/chain-workflow-demo.gif\" alt=\"Tech evaluation chain researching Zod via context7, producing a scored assessment table with security, performance, DX, integration, and ecosystem ratings\" width=\"720\" /\u003e\n\n\u003csub\u003eContext7 fetches live library docs mid-chain — final output is a structured assessment with sources\u003c/sub\u003e\n\n\u003c/details\u003e\n\nWhat happened:\n\n1. Loaded the `review` template with arguments\n2. Injected CAGEERF reasoning guidance\n3. Added a validation rule (AI self-evaluates against it)\n4. Chained output to the next step\n5. Ran a shell command for ground-truth validation\n6. Handed the final step off to a client-native subagent\n\n\u003e [!TIP]\n\u003e Chains support conditional branching, context threading, and agent handoffs.\n\u003e [Chains Lifecycle](docs/concepts/chains-lifecycle.md) · [MCP Tools Reference](docs/reference/mcp-tools.md)\n\n### Verification Loops\n\nGround-truth validation via shell commands — the AI keeps iterating until tests pass:\n\n```\n\u003e\u003eimplement-feature :: verify:\"npm test\" loop:true\n```\n\nImplements, runs the test, reads failures, fixes, retries. Spawns a fresh context after repeated failures to avoid context rot.\n\n| Preset      | Tries | Timeout | Use Case          |\n| ----------- | ----- | ------- | ----------------- |\n| `:fast`     | 1     | 30s     | Quick check       |\n| `:full`     | 5     | 5 min   | CI validation     |\n| `:extended` | 10    | 10 min  | Large test suites |\n\n\u003e [!TIP]\n\u003e **Autonomous test-fix cycles.** See [Ralph Loops](docs/guides/ralph-loops.md) for presets, timeout configuration, and context-rot prevention.\n\n### Judge Mode\n\nLet the AI pick the right resources for the task:\n\n```\n%judge Help me refactor this authentication module\n```\n\nAnalyzes available templates, reasoning frameworks, validation rules, and styles — applies the best combination automatically.\n\n\u003e [!TIP]\n\u003e **How judge mode selects resources.** See [Judge Mode Guide](docs/guides/judge-mode.md) for scoring, overrides, and preview with `%judge`.\n\n---\n\n## Run Anywhere\n\nAuthor workflows as YAML templates. Export as native skills to your client.\n\n```yaml\n# skills-sync.yaml — choose what to export\nregistrations:\n  claude-code:\n    user:\n      - prompt:development/review\n      - prompt:development/validate_work\n```\n\n```bash\nnpm run skills:export\n```\n\nThe `review` prompt becomes a `/review` Claude Code skill. `validate_work` becomes `/validate_work`. Same source, native experience — no MCP call required at runtime.\n\nCompiles to Claude Code skills, Cursor rules, OpenCode commands, and more. `npm run skills:diff` flags when exports drift from source.\n\n\u003cdetails\u003e\n\u003csummary\u003e\u003cstrong\u003eSee the export\u003c/strong\u003e — dry-run compile + skill preview\u003c/summary\u003e\n\n\u003cbr\u003e\n\n\u003cimg src=\"assets/demos/skills-export-demo.gif\" alt=\"Skills export dry-run compiling prompts to native skill files, then bat preview of the generated review skill with phases, gates, and arguments\" width=\"720\" /\u003e\n\n\u003csub\u003eDry-run compiles YAML templates into native client skills — review before writing\u003c/sub\u003e\n\n\u003c/details\u003e\n\n\u003e [!TIP]\n\u003e The [Skills Sync Guide](docs/guides/skills-sync.md) covers configuration, supported clients, and drift detection.\n\n---\n\n### With Hooks\n\nWell-composed prompts carry their own structure. Hooks keep the experience consistent across models and long sessions.\n\n\u003cdetails\u003e\n\u003csummary\u003eWhat hooks do\u003c/summary\u003e\n\nRoute operator syntax to the right tool automatically.\nTrack workflow progress across steps and long sessions.\nEnforce validation rules and step handoffs between agents.\n\n| Behavior            | What happens                                         |\n| ------------------- | ---------------------------------------------------- |\n| Prompt routing      | `\u003e\u003eanalyze` in conversation → correct MCP tool call  |\n| Chain continuity    | Injects step progress and continuation between steps |\n| Validation tracking | Tracks pass/fail verdicts across chain steps         |\n| Agent handoffs      | Routes `==\u003e` steps to client-native subagents        |\n| Session persistence | Preserves workflow state through context compaction  |\n\nHooks ship with the plugin install. Available for [Claude Code](.) (full), [OpenCode](https://github.com/minipuft/opencode-prompts) (full), [Gemini CLI](https://github.com/minipuft/gemini-prompts) (partial). Other clients: MCP tools only.\n\n→ [hooks/README.md](hooks/README.md)\n\n\u003c/details\u003e\n\n---\n\n\u003cdetails\u003e\n\u003csummary\u003e\u003cstrong\u003eSyntax Reference\u003c/strong\u003e\u003c/summary\u003e\n\n| Symbol | Name      | What It Does              | Example                |\n| :----: | :-------- | :------------------------ | :--------------------- |\n|  `\u003e\u003e`  | Prompt    | Execute template          | `\u003e\u003ecode_review`        |\n| `--\u003e`  | Chain     | Pipe to next step         | `step1 --\u003e step2`      |\n| `==\u003e`  | Handoff   | Route step to agent       | `step1 ==\u003e agent_step` |\n|  `*`   | Repeat    | Run prompt N times        | `\u003e\u003ebrainstorm * 5`     |\n|  `@`   | Framework | Inject reasoning guidance | `@CAGEERF`             |\n|  `::`  | Gate      | Add validation criteria   | `:: 'cite sources'`    |\n|  `%`   | Modifier  | Toggle behavior           | `%clean`, `%judge`     |\n|  `#`   | Style     | Apply formatting          | `#analytical`          |\n\n**Modifiers:**\n\n- `%clean` — No framework/gate injection\n- `%lean` — Gates only, skip framework\n- `%guided` — Force framework injection\n- `%judge` — AI selects best resources\n\n→ [MCP Tools Reference](docs/reference/mcp-tools.md) for full command documentation.\n\n\u003c/details\u003e\n\n\u003cdetails\u003e\n\u003csummary\u003e\u003cstrong\u003eThe Three Tools\u003c/strong\u003e\u003c/summary\u003e\n\n| Tool               | Purpose                                        |\n| ------------------ | ---------------------------------------------- |\n| `prompt_engine`    | Execute prompts with frameworks and validation |\n| `resource_manager` | Create, update, version, and export resources  |\n| `system_control`   | Status, analytics, framework switching         |\n\n```\nprompt_engine(command:\"@CAGEERF \u003e\u003eanalysis topic:'AI safety'\")\nresource_manager(resource_type:\"prompt\", action:\"list\")\nsystem_control(action:\"status\")\n```\n\n\u003c/details\u003e\n\n---\n\n## How It Works\n\n```mermaid\n%%{init: {'theme': 'neutral', 'themeVariables': {'background':'#0b1224','primaryColor':'#e2e8f0','primaryBorderColor':'#1f2937','primaryTextColor':'#0f172a','lineColor':'#94a3b8','fontFamily':'\"DM Sans\",\"Segoe UI\",sans-serif','fontSize':'14px','edgeLabelBackground':'#0b1224'}}}%%\nflowchart TB\n    classDef actor fill:#0f172a,stroke:#cbd5e1,stroke-width:1.5px,color:#f8fafc;\n    classDef server fill:#111827,stroke:#fbbf24,stroke-width:1.8px,color:#f8fafc;\n    classDef process fill:#e2e8f0,stroke:#1f2937,stroke-width:1.6px,color:#0f172a;\n    classDef client fill:#f4d0ff,stroke:#a855f7,stroke-width:1.6px,color:#2e1065;\n    classDef clientbg fill:#1a0a24,stroke:#a855f7,stroke-width:1.8px,color:#f8fafc;\n    classDef decision fill:#fef3c7,stroke:#f59e0b,stroke-width:1.6px,color:#78350f;\n\n    linkStyle default stroke:#94a3b8,stroke-width:2px\n\n    User[\"1. User sends command\"]:::actor\n    Example[\"\u003e\u003eanalyze @CAGEERF :: 'cite sources'\"]:::actor\n    User --\u003e Example --\u003e Parse\n\n    subgraph Server[\"MCP Server\"]\n        direction TB\n        Parse[\"2. Parse operators\"]:::process\n        Inject[\"3. Inject framework + gates\"]:::process\n        Render[\"4. Render prompt\"]:::process\n        Decide{\"6. Route verdict\"}:::decision\n        Parse --\u003e Inject --\u003e Render\n    end\n    Server:::server\n\n    subgraph Client[\"Claude (Client)\"]\n        direction TB\n        Execute[\"5. Run prompt + check gates\"]:::client\n    end\n    Client:::clientbg\n\n    Render --\u003e|\"Prompt with gate criteria\"| Execute\n    Execute --\u003e|\"Verdict + output\"| Decide\n\n    Decide --\u003e|\"PASS → render next step\"| Render\n    Decide --\u003e|\"FAIL → render retry prompt\"| Render\n    Decide --\u003e|\"Done\"| Result[\"7. Return to user\"]:::actor\n```\n\nCommand with operators → server parses and injects resources → client executes and self-evaluates → route: next step (pass), retry (fail), or return result (done).\n\n---\n\n## Documentation\n\n| I want to...                                             | Go here                                                                |\n| -------------------------------------------------------- | ---------------------------------------------------------------------- |\n| Build my first prompt                                    | [Prompt Authoring Tutorial](docs/tutorials/build-first-prompt.md)      |\n| Chain multi-step workflows                               | [Chains Lifecycle](docs/concepts/chains-lifecycle.md)                  |\n| Add validation to workflows                              | [Gates Guide](docs/guides/gates.md)                                    |\n| Use or create reasoning frameworks                       | [Methodologies Guide](docs/guides/methodologies.md)                    |\n| Use autonomous verification loops                        | [Ralph Loops](docs/guides/ralph-loops.md)                              |\n| Configure per-client MCP installs and `--client` presets | [Client Integration Guide](docs/guides/client-integration.md)          |\n| Compare client profile mapping and limitations           | [Client Capabilities Reference](docs/reference/client-capabilities.md) |\n| Export skills to other clients                           | [Skills Sync](docs/guides/skills-sync.md)                              |\n| Configure the server                                     | [CLI \u0026 Configuration](docs/guides/cli.md)                              |\n| Let the AI pick resources automatically                  | [Judge Mode Guide](docs/guides/judge-mode.md)                          |\n| Look up MCP tool parameters                              | [MCP Tools Reference](docs/reference/mcp-tools.md)                     |\n| Look up prompt YAML fields                               | [Prompt YAML Schema](docs/reference/prompt-yaml-schema.md)             |\n| Understand the architecture                              | [Architecture Overview](docs/architecture/overview.md)                 |\n| Fix common issues                                        | [Troubleshooting](docs/guides/troubleshooting.md)                      |\n\n---\n\n## Contributing\n\n```bash\ncd server\nnpm install\nnpm run build        # esbuild bundles to dist/index.js\nnpm test             # Run test suite\nnpm run validate:all # Full CI validation\n```\n\nThe build produces a self-contained bundle. `server/dist/` is gitignored — CI builds fresh from source.\n\nSee [CONTRIBUTING.md](CONTRIBUTING.md) for workflow details.\n\n---\n\n## License\n\n[AGPL-3.0](LICENSE)\n","project_url":"https://awesome.ecosyste.ms/api/v1/projects/github.com%2Fminipuft%2Fclaude-prompts","html_url":"https://awesome.ecosyste.ms/projects/github.com%2Fminipuft%2Fclaude-prompts","lists_url":"https://awesome.ecosyste.ms/api/v1/projects/github.com%2Fminipuft%2Fclaude-prompts/lists"}