By dynamic-dome
Suite of 7 skills, 7 slash commands, and 5 read-only checker hooks that turns rough ideas into agentic workflow blueprints, runs adversarial review-fix-verify loops, audits ecosystems as dynamic graphs, curates raw inbox material into wiki pages, runs MCP/DCO actions safely, and gates skill promotion. Built from the wiki/patterns/agentic-workflow-blueprints reference.
Execute an MCP/external-tool action with privacy gate, output validation, and audit trail
Run Review-Fix-Verify on a code change with separated phases and evidence requirement
Check whether a recurring pattern should become a real Skill (promote / keep-as-pattern / already-covered / reject)
Promote a raw inbox file (image, voice transcript, telegram note) into a curated wiki page
Run a dynamic-graph audit on a broad scope (plugin ecosystem, wiki cluster, repo monorepo)
Use when the user wants to turn a rough idea, screenshot, diagram, voice note, raw inbox file, or hand-drawn sketch into a concrete agentic workflow with triggers, inputs, steps, verification, error paths, and a Skill/Command/Hook/MCP recommendation. Trigger phrases - "design a workflow", "turn this into a skill", "make a blueprint", "spezifiziere", "mach daraus ein workflow", "skill blueprint", "/workflow design". Do NOT use this skill to execute workflows or write production code; the output is a blueprint, not an implementation. Also skip when the user only wants a diagram and is not asking for a runnable agentic process.
Use when work routes through DCO (Dynamic Central Orchestrator), Telegram bot, Mini-App, or MCP tools and crosses privacy/data classification boundaries. Trigger phrases - "dco safe run", "mcp action with validation", "telegram tool gate", "/mcp safe-run", "mit tool grenzen", "privacy gate". Validates tool output before persisting to DB/wiki/answer. Skip for purely local file edits without external tools and skip when no privacy class change is involved.
Use when a broad ecosystem (plugin landscape, wiki cluster, repo monorepo, DCO function map) needs an audit where the actual graph emerges through exploration, not from a pre-defined plan. Trigger phrases - "audit the plugin ecosystem", "schau dir alles an", "inventur machen", "find redundancies", "map the wiki", "/workflow audit", "ecosystem review". Returns inventory, clusters, role/dataflow map, redundancies, synergies, prioritized TODOs. Skip when the scope is small enough for a single-loop pass or when the audit was just done within the last 7 days.
Use when raw inbox material (image, voice transcript, telegram note, link) in wiki/raw/ should be promoted into a curated wiki page with proper frontmatter, sources, and cross-links. Trigger phrases - "curate this raw file", "promote raw to wiki", "kuratiere die raw datei", "ins wiki ueberfuehren", "raw inbox aufraeumen", "/wiki curate". The skill writes a NEW wiki page; the raw file stays untouched. Skip when the user wants only to read raw content without producing a wiki artifact, or when the source is not in wiki/raw/.
Use when a code change is risky enough that a linear patch is not safe — separates Review (find findings), Fix (apply minimal patch), and Verify (run tests/lints/UI checks) into distinct phases with feedback. Trigger phrases - "review-fix-verify", "review this patch", "kritischer review", "pruef das nochmal", "patch reviewen", "/review-loop". Each finding requires evidence (file:line, test, screenshot). Skip when the change is trivial or already verified, and skip when there is no test/verification surface (then prefer plain code-reviewer).
Modifies files
Hook triggers on file write and edit operations
Own this plugin?
Verify ownership to unlock analytics, metadata editing, and a verified badge. GitHub access is read-only (username + org membership).
Sign in to claimOwn this plugin?
Verify ownership to unlock analytics, metadata editing, and a verified badge. GitHub access is read-only (username + org membership).
Sign in to claimBased on adoption, maintenance, documentation, and repository signals. Not a security audit or endorsement.
A Claude Code plugin that turns rough ideas, raw inbox material, and ad-hoc workflows into a coherent operating system for agentic work — with 7 skills, 7 slash commands, and 5 read-only checker hooks.
Built from wiki/patterns/agentic-workflow-blueprints.md. Tested through 3 real designer runs (TODO file, concept page, multi-phase roadmap) before being expanded into a full plugin.
| Layer | Bausteine |
|---|---|
| Skills (7) | agentic-workflow-designer, workflow-router, raw-inbox-curator, dynamic-graph-auditor, review-fix-verify-loop, dco-tool-safe-runner, skill-promotion-checker |
| Slash Commands (7) | /workflow design, /workflow route, /workflow audit, /wiki curate, /review-loop, /skill promote, /mcp safe-run |
| Hooks (5, read-only) | raw-write-guard, wiki-curation-check, mcp-output-envelope, private-export-gate, workflow-wrap-up |
| Skill | Zweck | Output |
|---|---|---|
agentic-workflow-designer | Quelle -> Workflow-Blueprint mit Bauteil-Empfehlung | Markdown-Block mit Schema |
workflow-router | Anfrage -> Archetyp-Routing (A-G) | YAML mit archetype, risk, next_skill |
raw-inbox-curator | Raw-Datei -> kuratierte Wiki-Seite | Neue Wiki-Seite + Index/Log-Updates |
dynamic-graph-auditor | Scope -> Inventar + Cluster + Redundanzen + TODOs | YAML-Audit-Report + optionaler Mermaid |
review-fix-verify-loop | Patch -> Findings -> Fixes -> Verifikation | YAML mit getrennten Phasen |
dco-tool-safe-runner | MCP-/Tool-Aktion mit Privacy-Gate + Validierung | YAML mit Audit-Trail |
skill-promotion-checker | Pattern -> Promotion-Verdict (4 Werte) | YAML mit verdict, 6 Checks, Naming-Collision |
Alle Hooks warnen, blocken nicht. Sie schreiben Audit-Logs unter ~/.claude/logs/<hook-name>.ndjson. Promotion-Pfad zu echten Block-Hooks erst nach 10+ false-positive-freien Laeufen — siehe skills/agentic-workflow-designer/references/hooks-readonly.md.
| Hook | Event | Was er prueft |
|---|---|---|
raw-write-guard | PreToolUse (Write/Edit/...) | Ziel ist nicht in wiki/raw/ |
wiki-curation-check | PostToolUse (Wiki-Write) | Frontmatter, Index-Link, Log-Eintrag |
mcp-output-envelope | PostToolUse (mcp__*) | MCP-Output als untrusted markieren |
private-export-gate | PreToolUse (Export-Tools) | Privacy-Klasse der Quelle |
workflow-wrap-up | Stop | Reminder fuer Wrap-up |
claude plugin marketplace add willneverusegit/agentic-workflow-suite
claude plugin install agentic-workflow-suite
# In einem beliebigen Projekt-Root:
git clone https://github.com/willneverusegit/agentic-workflow-suite.git
# Plugin von einem lokalen Pfad referenzieren:
claude plugin install agentic-workflow-suite --from ./agentic-workflow-suite
ln -s "$(pwd)/skills/agentic-workflow-designer" ~/.claude/skills/agentic-workflow-designer
Hooks sind opt-in. Konfiguration in hooks/hooks.json einsehen, dann gewuenschte Eintraege in deine ~/.claude/settings.json (oder Projekt-.claude/settings.json) uebernehmen — nicht alle muessen aktiviert sein.
.claude-plugin/
└── plugin.json Plugin-Manifest
skills/
├── agentic-workflow-designer/ Hauptdesigner mit references/ + examples/
├── workflow-router/
├── raw-inbox-curator/
├── dynamic-graph-auditor/
├── review-fix-verify-loop/
├── dco-tool-safe-runner/
└── skill-promotion-checker/
commands/
├── workflow-design.md /workflow design <source>
├── workflow-route.md /workflow route <request>
├── workflow-audit.md /workflow audit <scope>
├── wiki-curate.md /wiki curate <raw-pfad>
├── review-loop.md /review-loop <scope>
├── skill-promote.md /skill promote <pattern>
└── mcp-safe-run.md /mcp safe-run <tool+intent>
hooks/
├── hooks.json Settings-Snippet
├── raw-write-guard.py
├── wiki-curation-check.py
├── mcp-output-envelope.py
├── private-export-gate.py
└── workflow-wrap-up.py
Der Hauptdesigner durchlief vor dem Plugin-Bundle drei dokumentierte Realitaetslaufe:
npx claudepluginhub dynamic-dome/claude-plugins-marketplace --plugin agentic-workflow-suiteAdapter-agnostischer 4-Rollen-Loop (Orchestrator/Researcher/Builder/Judge) mit persistierenden State-Dateien pro Projekt.
Autonomous improvement loop for Claude Code projects — decomposes goals into tasks, executes them via Worker/Judge subagents, mines session logs for error patterns, and generates evidence-based CLAUDE.md improvement suggestions.
Interactive Claude Code workshop plugin — two modes: /workshop guide (moderator co-pilot) and /workshop learn (self-paced tutor). 17 modules (5+5+7) across 3 blocks covering foundations, ecosystem, and advanced topics.
Autonomer Meta-Agent-Orchestrator: Ein Opus 4.6 Agent der Haiku-Brainstormer, NotebookLM-RAG und Codex-Instanzen orchestriert um komplexe Aufgaben autonom zu loesen.
Interactive Claude Code workshop plugin — two modes: /workshop guide (moderator co-pilot) and /workshop learn (self-paced tutor). 13 modules across 3 blocks covering foundations, ecosystem, and advanced topics.
Harness-native ECC operator layer - 67 agents, 271 skills, 92 legacy command shims, reusable hooks, rules, selective install profiles, and production-ready workflows for Claude Code, Codex, OpenCode, Cursor, and related agent harnesses
Superpowers Plus core skills library for Claude Code: planning, execution routing, TDD, debugging, and collaboration workflows
Comprehensive C4 architecture documentation workflow with bottom-up code analysis, component synthesis, container mapping, and context diagram generation
Core skills library for Claude Code: TDD, debugging, collaboration patterns, and proven techniques
Tools to maintain and improve CLAUDE.md files - audit quality, capture session learnings, and keep project memory current.
Upstash Context7 MCP server for up-to-date documentation lookup. Pull version-specific documentation and code examples directly from source repositories into your LLM context.