Getting it into your agent
One page per mod, every tool's command on it. A separate URL per tool would split the same page into five that compete with each other.
npx skills add Momo2323-ui/claude-orchestra --skill orchestra-routergit clone --depth 1 https://github.com/Momo2323-ui/claude-orchestraWrote this? Show the measurements
A badge with what this costs and how it scanned, read live from this page, so it follows the numbers instead of freezing them. Markdown for a README, HTML for a documentation site or a project page.
[](https://agentmods.dev/skills/momo2323-ui/claude-orchestra/orchestra-router)<a href="https://agentmods.dev/skills/momo2323-ui/claude-orchestra/orchestra-router"><img src="https://agentmods.dev/badge/skills/momo2323-ui/claude-orchestra/orchestra-router/github.svg" alt="Measured on agentmods" height="20"></a>Or the 80×15 button, for a site that already has a row of RSS and ATOM ones. Only the verdict fits; the numbers stay here.
<a href="https://agentmods.dev/skills/momo2323-ui/claude-orchestra/orchestra-router"><img src="https://agentmods.dev/badge/skills/momo2323-ui/claude-orchestra/orchestra-router.svg" alt="Reviewed on agentmods" width="80" height="20"></a>What it costs to keep this loaded
Counted locally with the o200k_base tokenizer, which is exact for GPT models; Claude uses its own tokenizer and its counts differ. Treat this as one consistent yardstick across the catalogue rather than a bill. Prices are per million input tokens.
| Model | Per session | Once invoked |
|---|---|---|
| Fable 5.1 | $0.00110 | $0.02524 |
| Opus 5 | $0.00055 | $0.01262 |
| Sonnet 5 | $0.00022 | $0.00505 |
| Haiku 4.5 | $0.00011 | $0.00252 |
Grade A, and why
orchestra-router scanned grade A with 0 findings against 26 rules in 11 categories — prompt injection, anti-refusal, data exfiltration, privilege escalation, supply chain, agent snooping, system-prompt leakage, SSRF and excessive agency — measured 11d ago.
A static scan of the body, not an audit. Every finding is printed with the line that produced it so you can judge whether it matters here. A mod is markdown that instructs an agent; that is exactly why what it instructs is worth reading.
Nothing flagged
None of the 26 patterns this scan looks for appear in this file: no shell pipes, no recursive deletes, no credential paths, no hidden text, no instruction-override or anti-refusal phrasing, no agent-config snooping. That is not a guarantee, it is the absence of the things that are checkable.
How it starts
The opening of the file, as written. The whole thing — 154 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Orchestra Router
Routes requests to the right orchestra(s). v3 doctrine: route silently — announce only delegation,
in one line. This is the runtime companion to the reference doctrine at
~/.claude/rules/orchestra-system.md.
Routing algorithm (run on every task)
-
Check NEXUS first. If the message signals ideation or business planning ("I have an idea", "thinking of building", "plan a business", "what if I built", "is this a good idea", "I've been thinking about", "should I build") → invoke the
nexus-strategyskill. NEXUS becomes the meta-conductor and sequences orchestras across its 7 phases. This overrides single-orchestra routing — err toward firing it. -
Match intent to orchestra(s). Use the trigger table below. Match on meaning, not just exact words. A request can match several — that's a stack.
-
Detect stacks. Compound requests activate multiple orchestras in sequence. Examples:
- "make a landing page" → PLANNING → DESIGN → BUILD (QA inside) → optionally SEO/GROWTH
- "launch my app" → MARKETING → CONTENT → SEO → PAID ADS → GROWTH → ANALYTICS
- "find clients and email them" → LEAD GEN → CONTENT (email) → (ASK before sending)
-
Announce delegation only, in one line. No ensemble blocks. Format:
→ <conductor/agent>: <task>(e.g.→ search-agent: competitor pricing sweep). Routing that doesn't delegate stays silent — the work leads the reply. -
Let the conductor lead. Hand the work to the orchestra's conductor agent, which sequences its First Chair + Section players. One conductor per orchestra — no free-for-all.
-
Respect allowances + gates. Honor each orchestra's ASK-before rules and quality gate (see constitution). Reserve Bench agents never auto-fire — explicit name only.
Trigger → Orchestra quick table
| If the message is about… | Activate |
|---|---|
| ideation, business idea, "what if", "should I build" | NEXUS (meta) → orchestras |
| build, fix, refactor, implement, deploy, code review, hard bug, diagnose, flaky test, can't reproduce, intermittent bug, investigate this error, debug this | 🏗️ BUILD |
| UI, UX, design, mockup, Figma, layout, brand visuals, component, 21st.dev, design inspiration, "give me UI ideas", build a component | 🎨 DESIGN — reach for Magic / 21st.dev MCP FIRST (component inspiration + generation) |
| research, investigate, competitive analysis, deep dive | 🔍 RESEARCH |
| marketing strategy, social, campaign, GTM, awareness | 📣 MARKETING |
| copy, blog, newsletter, carousel, email sequence | ✍️ CONTENT |
| SEO, rank, schema, AI search, GEO, AEO | 📈 SEO+GEO |
| find leads/clients, prospect, outreach, sales pipeline | 💼 LEAD GEN |
| PRD, product strategy, roadmap, prioritization | 🧩 PRODUCT |
| video, reels, image gen, thumbnail, Higgsfield | 🎬 VIDEO+MEDIA |
| analytics, metrics, funnel, cohort, A/B results, demand forecasting, trend prediction | 📊 ANALYTICS |
| remember, knowledge graph, Obsidian, map codebase, query NotebookLM, ask my notebooks (Python), notebooklm-scripts | 🧠 KNOWLEDGE |
| report, deck, PDF, doc, slides, save to Notion/Drive | 📄 DOCUMENTS |
| Google/Meta ads, PPC, paid media, ad creative | 💰 PAID ADS |
| automate, workflow, n8n, "when X do Y", recurring | ⚙️ AUTOMATION |
| stealth browser, cloakbrowser, bypass bot detection, bypass Cloudflare, bypass DataDome, fingerprint evasion, scrape without detection, AI agent browsing undetected | ⚙️ AUTOMATION |
| AI pipeline, multi-agent, build MCP, RAG, fine-tune, time-series forecast, TimesFM, predict future values, forecasting model | 🤖 AI/ML |
| iOS, Swift, SwiftUI, Xcode | 📱 iOS/SWIFT |
| plan, sprint, OKRs, user stories, pre-mortem | 📋 PLANNING |
| conversion, churn, paywall, ASO, pricing, retention | 🚀 GROWTH |
| invest, stock, valuation, DCF, earnings, portfolio | 💵 PERSONAL FINANCE |
| as CEO/CFO/CTO, founder advice, board deck, M&A, scenario plan, org health, strategic decision, devil's advocate, stress-test this | 👔 EXECUTIVE ADVISORY |
| RevOps, deal desk, pricing strategy, partnerships, RFP, customer success | 💼 LEAD GEN & SALES (RevOps section) |
| process mapping, vendor mgmt, procurement, capacity planning | ⚙️ AUTOMATION & OPS (Business Ops) |
| security incident, pentest, malware, forensics, DFIR, threat intel, CTI, vulnerability, CVE, exploit, ransomware, phishing investigation, red team, blue team, purple team, zero trust, OSINT, IOC, TTPs, MITRE ATT&CK, compliance audit, NIST, ISO 27001, PCI-DSS, OT/ICS/SCADA security, cloud breach, lateral movement, C2, threat hunting, detection engineering, Sigma, Volatility, Ghidra, BloodHound | 🛡️ CYBERSECURITY |
What this file has done since we first saw it
Hashed on every crawl. A supply-chain change to an agent config is a question of when, not whether, so the history is kept rather than the latest state alone.
- 11d ago First seen · 154 lines · 110 tokens per session scan A d0af97bad5b1
orchestra-router is a skill published in the GitHub repository Momo2323-ui/claude-orchestra (8 stars, last pushed 1mo ago), licensed MIT. It adds 110 tokens to every session and 2,524 once invoked, about $0.0006 per session on Opus 5. A static security scan graded it A with 0 findings. No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-31.
Other skills, from other repositories
mle-workflow
Production machine-learning engineering workflow for data contracts, reproducible training, model evaluation, deployment, monitoring, and rollback. Use when building, reviewing, or hardening ML systems beyond one-off notebooks.
autonomous-loops
Patterns and architectures for autonomous Claude Code loops — from simple sequential pipelines to RFC-driven multi-agent DAG systems.
react-patterns
React 18/19 patterns including hooks discipline, server/client component boundaries, Suspense + error boundaries, form actions, data fetching, state management decision trees, and accessibility-first composition. Use when writing or reviewing React components.
article-writing
Write articles, guides, blog posts, tutorials, newsletter issues, and other long-form content in a distinctive voice derived from supplied examples or brand guidance. Use when the user wants polished written content longer than a paragraph, especially when voice consistency, structure, and credibility matter.
security-review
Perform a focused security review of pending git changes to identify high-confidence security vulnerabilities with real exploitation potential. Use this skill when the user asks for a security review, security audit, vulnerability scan, or wants to check pending changes on a branch for security issues before merging.…
huggingface-llm-trainer
Train or fine-tune language models with TRL or Unsloth on Hugging Face Jobs, including SFT, DPO, GRPO, reward models, and GGUF conversion. Use for cloud LLM training; use huggingface-vision-trainer for vision tasks.