orchestrate
Plan and execute multi-agent workflows through lionagi's orchestration MCP server. The
plugin ships one MCP tool, mcp__plugin_orchestrate_lion__request, which dispatches
namespaced verbs behind a single ops array — not a CLI, not one tool per verb.
Start here, every session
mcp__plugin_orchestrate_lion__request(help=true)
Call this once, in its own call (never combined with ops), before writing any spawn call.
It returns the verb catalog, each verb's required parameters, and either its
schema_fingerprint or a marker that the fingerprint varies by playbook (see
reference.md). Never guess a parameter name — this call and targeted help are the source
of truth.
When to use which verb
| Situation |
Verb |
| Single task, one agent |
agent.submit |
| Same prompt to N independent workers |
fanout.submit |
| Staged pipeline with dependencies |
flow.submit |
| Pre-saved parametric workflow |
play.submit |
| Check on a run already submitted |
job.status, job.wait, job.output |
If all subtasks are independent (no output feeds another), use fanout.submit.
If any subtask depends on the output of another, use flow.submit.
play.submit loads a saved prompt and defaults, then uses the same planner and executor as
flow.submit.
Every *.submit call spawns a detached background run and returns a run id right
away — there is no blocking call that returns a finished answer. Read the result back with
job.status, job.wait, or job.output.
Quick start
# Catalog + fingerprints, once per session
mcp__plugin_orchestrate_lion__request(help=true)
# Single agent
mcp__plugin_orchestrate_lion__request(ops=[
{"op": "agent.submit", "args": {"query": ["claude"], "prompt": "Write unit tests for auth.py"},
"schema_fingerprint": "<from help>"}
])
# Fan out 4 parallel workers + synthesize
mcp__plugin_orchestrate_lion__request(ops=[
{"op": "fanout.submit",
"args": {"query": ["claude"], "prompt": "Review this codebase for security issues", "num_workers": 4, "with_synthesis": true},
"schema_fingerprint": "<from help>"}
])
# DAG flow — dry-run first, then commit
mcp__plugin_orchestrate_lion__request(ops=[
{"op": "flow.submit",
"args": {"query": ["claude"], "prompt": "Audit auth, implement fixes, verify with tests", "dry_run": true},
"schema_fingerprint": "<from help>"}
])
# ... inspect the plan, then resend without dry_run ...
# Run a saved playbook
mcp__plugin_orchestrate_lion__request(ops=[
{"op": "play.submit", "args": {"playbook": "feature", "prompt": "Add JWT middleware"},
"schema_fingerprint": "<from help={'verb': 'play.submit', 'playbook': 'feature'}>"}
])
# Check a run
mcp__plugin_orchestrate_lion__request(ops=[
{"op": "job.wait", "args": {"run_ids": ["<run-id>"], "max_wait": 60}}
])
Key principles
help=true in its own call, before writing a spawn call — never combine help and
ops in one call; the server refuses it.
schema_fingerprint rides as a sibling of args, never inside it — every
*.submit verb requires it, fetched from help.
- Named profiles are installation-specific — call
profile.list before passing agent,
or pass a model through the verb's positional query argument where supported.
- Critic assignments depend on producers — a critic is an ordinary
TaskAssignment, not
a special control node. Put it after every output it reviews.
- Every assignment gets a separate branch — repeated roles do not share conversation
memory. Pass results through declared dependencies and artifacts.
- Artifact paths are supplied by the engine — name outputs clearly and use the exact
upstream directories injected into downstream assignment context; do not guess
../ paths.
- Use
depends_on only for real data dependencies — independent assignments leave it
empty and run concurrently.
- Reserve reactive capacity —
max_ops is shared by the initial plan and follow-up spawns;
a plan that fills it leaves no room for reactive work.
dry_run is a text preview — it does not build a run graph. show_graph writes a graph
artifact only after an executing flow finishes.
- The catalog is the authority, not this bundle —
reference.md documents the verbs
these skills call, which is a subset of what the server offers; and the shipped server
resolves the latest lionagi release, so a verb you found by reading a checkout may not
be there yet. help=true settles both questions.
Companion references
For detailed documentation, read these companion files in this skill directory:
- reference.md — the MCP tool's parameter shapes for the verbs these
skills call, the
schema_fingerprint protocol, and the secondary li CLI flag tables
- dag-planning.md —
TaskAssignment data model, dependency planning, branch and artifact handoff, reactive budgets
- workflows.md — standard workflow patterns (parallel exploration, staged pipeline, long-running work, playbooks, visualization)
- teams-and-tracking.md — team coordination patterns, invocation tracking, scheduling — and which of these are CLI-only today
Source code
For someone working inside a lionagi checkout — informational, not part of the MCP surface:
| Component |
Path |
| MCP server + verb catalog |
lionhub/src/lionhub/serve/mcp/server.py, lionhub/src/lionhub/serve/mcp/verbs.py, lionhub/src/lionhub/serve/mcp/dispatch.py |
| CLI entrypoint |
lionhub/src/lionhub/serve/cli/main.py |
Flow engine (TaskAssignment DAG) |
lionhub/src/lionhub/serve/cli/orchestrate/flow.py |
| Fanout engine |
lionhub/src/lionhub/serve/cli/orchestrate/fanout.py |
| Argparse definitions |
lionhub/src/lionhub/serve/cli/orchestrate/__init__.py |
| Agent CLI |
lionhub/src/lionhub/serve/cli/agent.py |
| Teams |
lionhub/src/lionhub/serve/cli/team.py |
| Invocations |
lionhub/src/lionhub/serve/cli/invoke.py |
| Scheduler engine |
lionhub/src/lionhub/orchestration/scheduler/engine.py |
1---2name: orchestrate3description: >4 Plan and execute multi-agent workflows through lionagi's orchestration MCP tool:5 flow.submit (DAG pipelines), fanout.submit (parallel workers), agent.submit (one6 agent), and play.submit (saved playbooks). Use when a task needs multiple agents7 working in parallel or staged phases.8allowed-tools: [Bash, Read, Write, Glob, Grep]9metadata:10 source: https://skillmd.com/skills/ohdearquant/orchestrate11---1213# orchestrate1415Plan and execute multi-agent workflows through lionagi's orchestration MCP server. The16plugin ships one MCP tool, `mcp__plugin_orchestrate_lion__request`, which dispatches17namespaced verbs behind a single `ops` array — not a CLI, not one tool per verb.1819## Start here, every session2021```22mcp__plugin_orchestrate_lion__request(help=true)23```2425Call this once, in its own call (never combined with `ops`), before writing any spawn call.26It returns the verb catalog, each verb's required parameters, and either its27`schema_fingerprint` or a marker that the fingerprint varies by playbook (see28`reference.md`). Never guess a parameter name — this call and targeted `help` are the source29of truth.3031## When to use which verb3233| Situation | Verb |34|---|---|35| Single task, one agent | `agent.submit` |36| Same prompt to N independent workers | `fanout.submit` |37| Staged pipeline with dependencies | `flow.submit` |38| Pre-saved parametric workflow | `play.submit` |39| Check on a run already submitted | `job.status`, `job.wait`, `job.output` |4041If all subtasks are independent (no output feeds another), use `fanout.submit`.42If any subtask depends on the output of another, use `flow.submit`.43`play.submit` loads a saved prompt and defaults, then uses the same planner and executor as44`flow.submit`.4546Every `*.submit` call spawns a **detached background run** and returns a run id right47away — there is no blocking call that returns a finished answer. Read the result back with48`job.status`, `job.wait`, or `job.output`.4950## Quick start5152```53# Catalog + fingerprints, once per session54mcp__plugin_orchestrate_lion__request(help=true)5556# Single agent57mcp__plugin_orchestrate_lion__request(ops=[58 {"op": "agent.submit", "args": {"query": ["claude"], "prompt": "Write unit tests for auth.py"},59 "schema_fingerprint": "<from help>"}60])6162# Fan out 4 parallel workers + synthesize63mcp__plugin_orchestrate_lion__request(ops=[64 {"op": "fanout.submit",65 "args": {"query": ["claude"], "prompt": "Review this codebase for security issues", "num_workers": 4, "with_synthesis": true},66 "schema_fingerprint": "<from help>"}67])6869# DAG flow — dry-run first, then commit70mcp__plugin_orchestrate_lion__request(ops=[71 {"op": "flow.submit",72 "args": {"query": ["claude"], "prompt": "Audit auth, implement fixes, verify with tests", "dry_run": true},73 "schema_fingerprint": "<from help>"}74])75# ... inspect the plan, then resend without dry_run ...7677# Run a saved playbook78mcp__plugin_orchestrate_lion__request(ops=[79 {"op": "play.submit", "args": {"playbook": "feature", "prompt": "Add JWT middleware"},80 "schema_fingerprint": "<from help={'verb': 'play.submit', 'playbook': 'feature'}>"}81])8283# Check a run84mcp__plugin_orchestrate_lion__request(ops=[85 {"op": "job.wait", "args": {"run_ids": ["<run-id>"], "max_wait": 60}}86])87```8889## Key principles9091- **`help=true` in its own call, before writing a spawn call** — never combine `help` and92 `ops` in one call; the server refuses it.93- **`schema_fingerprint` rides as a sibling of `args`**, never inside it — every94 `*.submit` verb requires it, fetched from `help`.95- **Named profiles are installation-specific** — call `profile.list` before passing `agent`,96 or pass a model through the verb's positional `query` argument where supported.97- **Critic assignments depend on producers** — a critic is an ordinary `TaskAssignment`, not98 a special control node. Put it after every output it reviews.99- **Every assignment gets a separate branch** — repeated roles do not share conversation100 memory. Pass results through declared dependencies and artifacts.101- **Artifact paths are supplied by the engine** — name outputs clearly and use the exact102 upstream directories injected into downstream assignment context; do not guess `../` paths.103- **Use `depends_on` only for real data dependencies** — independent assignments leave it104 empty and run concurrently.105- **Reserve reactive capacity** — `max_ops` is shared by the initial plan and follow-up spawns;106 a plan that fills it leaves no room for reactive work.107- **`dry_run` is a text preview** — it does not build a run graph. `show_graph` writes a graph108 artifact only after an executing flow finishes.109- **The catalog is the authority, not this bundle** — `reference.md` documents the verbs110 these skills call, which is a subset of what the server offers; and the shipped server111 resolves the latest lionagi *release*, so a verb you found by reading a checkout may not112 be there yet. `help=true` settles both questions.113114## Companion references115116For detailed documentation, read these companion files in this skill directory:117118- **[reference.md](reference.md)** — the MCP tool's parameter shapes for the verbs these119 skills call, the `schema_fingerprint` protocol, and the secondary `li` CLI flag tables120- **[dag-planning.md](dag-planning.md)** — `TaskAssignment` data model, dependency planning, branch and artifact handoff, reactive budgets121- **[workflows.md](workflows.md)** — standard workflow patterns (parallel exploration, staged pipeline, long-running work, playbooks, visualization)122- **[teams-and-tracking.md](teams-and-tracking.md)** — team coordination patterns, invocation tracking, scheduling — and which of these are CLI-only today123124## Source code125126For someone working inside a lionagi checkout — informational, not part of the MCP surface:127128| Component | Path |129|---|---|130| MCP server + verb catalog | `lionhub/src/lionhub/serve/mcp/server.py`, `lionhub/src/lionhub/serve/mcp/verbs.py`, `lionhub/src/lionhub/serve/mcp/dispatch.py` |131| CLI entrypoint | `lionhub/src/lionhub/serve/cli/main.py` |132| Flow engine (`TaskAssignment` DAG) | `lionhub/src/lionhub/serve/cli/orchestrate/flow.py` |133| Fanout engine | `lionhub/src/lionhub/serve/cli/orchestrate/fanout.py` |134| Argparse definitions | `lionhub/src/lionhub/serve/cli/orchestrate/__init__.py` |135| Agent CLI | `lionhub/src/lionhub/serve/cli/agent.py` |136| Teams | `lionhub/src/lionhub/serve/cli/team.py` |137| Invocations | `lionhub/src/lionhub/serve/cli/invoke.py` |138| Scheduler engine | `lionhub/src/lionhub/orchestration/scheduler/engine.py` |