- Home
- Skills
- Agents & Orchestration
- Master Orchestrator
Works with the AI Agents you already use
Master Orchestrator
Route ambiguous engineering work to the single smallest-sufficient next phase: at most one primary specialist, verified capabilities, and an explicit stop condition.
$15
Master Orchestrator
Example session with this skill installed
Route this synthetic task using the package's route output contract: "Our checkout sometimes double-charges. Figure it out and fix it." In the example context, shell and Git are available and browser access is absent. Identify only the next bounded phase, authority, mutation level, capability assumptions, and stop condition. Do not inspect a repository, run tools, claim findings, or execute a fix.
- Read your context and instructions
- Compiled the master orchestrator
Illustrative example based on the synthetic worked example included in the package. No repository was inspected, no shell or Git command was run, and no checkout finding is asserted for this listing.
MODE: FIX
PRIMARY: root-cause-debugging
VALIDATOR: none yet (no diff exists to review)
AUTHORITY: repo read; no deploy.
MUTATION LEVEL: read-only until cause is established.
CAPABILITY ASSUMPTIONS: shell and Git are available in this example; browser is absent, so any rendered check is UNVERIFIED.
WHY: The sample prompt reports incorrect behavior but supplies no established cause or reproduction; diagnosis precedes editing. Double charging makes idempotency or retry behavior a hypothesis to investigate, not a proven cause.
STOP CONDITION: Stop after a causal hypothesis has a red-capable reproduction; then route the fix to bounded-implementation and review to change-assurance-review after a concrete diff exists.
Connects securely to your tools. The creator never sees your data.
About this skill
Master Orchestrator is a read-only router for ambiguous, multi-step, or risky engineering work. It selects the single smallest-sufficient next phase, at most one primary specialist, and at most one validator only when a concrete object already exists. It checks required capabilities and names an explicit stop condition; downstream execution remains with the selected workflow.
When to use it
Use it when the request spans multiple phases, its owner is unclear, the cause of a failure is unknown, or the work has a meaningful risk boundary. Skip it for a trivial bounded task whose owner and authority are already obvious.
How it works
- Establish the requested outcome, current phase, highest project authority, mutation level, and evidence available.
- Read the smallest sufficient project evidence and verify the capabilities needed for the next phase.
- Select one routing mode: FAST, EXPLORE, DESIGN, IMPLEMENT, FIX, REVIEW, UI, CONTINUE, or SHIP.
- Choose at most one primary specialist. Add one validator only when a concrete object already exists for validation. If a required specialist or capability is unavailable, name the process or fallback instead of pretending it ran.
- Return the route record and stop at the handoff.
Route record
MODE: PRIMARY: VALIDATOR: AUTHORITY: MUTATION LEVEL: CAPABILITY ASSUMPTIONS: WHY: STOP CONDITION:
Boundaries
Project governance, accepted decisions, approved specs, committed contracts, tests, code, and current evidence outrank the skill. The router does not execute the selected workflow, create a fix, or certify a downstream result. It keeps one writer per repository surface, prevents recursive routing without changed state, and escalates rather than improvising across security, irreversible data, migrations, payments, production deployment, or unresolved architecture authority. Installs, commits, pushes, merges, deploys, and destructive migrations are never implied.
How it differs from a downstream specialist
A specialist owns its bounded phase, such as diagnosing an unknown failure or building a throwaway prototype. Master Orchestrator chooses the next phase and handoff boundary; it stops before that specialist's work begins.
Find more skills like this on the Claude Skills Marketplace, the Codex Skills Marketplace and the Gemini Skills Marketplace, or browse Claude Skills for Agents & Orchestration.
The route depends on an accurate task objective, available project authority, and enough current evidence to identify the next phase. It selects only the next bounded phase, not a complete multi-phase execution plan. It does not run the selected workflow or validate downstream results; unavailable evidence or capabilities remain UNVERIFIED or BLOCKED unless a supported fallback answers the routing question.
How to use
- 1
Connect your agent to the Agensi MCP
Copy the prompt and paste it into Claude, ChatGPT, Cursor or any agent that supports MCP. It sets up the connection for you. Manual setup
- 2
Sign in to Agensi
Your agent shows a sign-in link. Sign in and click Allow.
- 3
Ask for what you need
Your agent loads this skill and uses it on your task. It always asks before buying a skill.
Reviews
No reviews yet
Be one of the first to try it. Every listed skill goes through the trust checks below.
Security scanned
Checked by our 8-point scan before listing
Fresh listing
Recently published to Agensi
Money-back guarantee
Doesn't do what its page promises? You get your money back.
Trust & safety
Security scan
Security scan passed 17 days ago
Permissions
Read Files is required to inspect project authority and the smallest sufficient current evidence. Terminal/Shell is an optional enhancement for targeted search or verifying whether a requested runner is available; use supplied files or host-native discovery when possible. Browser, Network Access, Write Files, MCP, external accounts, and Environment Variables are not required by this read-only router. The selected downstream phase may have separate capability requirements; verify them before handoff and mark UNVERIFIED or BLOCKED if no evidence-supported fallback exists. Installs, commits, pushes, merges, deploys, and destructive migrations are not implied.