New: 98,000+ skills indexed from GitHub repos with 100+ stars · every skill safety-checkedBrowse top skillsTop skills
Esc to closeBrowse all skills
Submit a skillBrowse skills
Submit a skillBrowse skills

Agent skill

Orchestration

by xiaolai in xiaolai/nlpm

Multi-agent workflow patterns: parallel dispatch, pipelines, QC gates, retries.

ISCAuto-check passedAgent Workflows

Install Orchestration

skills CLI
$ npx skills add xiaolai/nlpm --skill orchestration -a claude-code

Project install by default; add -g for ~/.claude/skills/.

GitHub CLI
$ gh skill install xiaolai/nlpm orchestration --agent claude-code

Project scope by default; add --scope user for a personal install. Needs GitHub CLI 2.90.0 or later (public preview).

Manual copy
$ git clone --depth 1 https://github.com/xiaolai/nlpm.git skills-src && mkdir -p .claude/skills && cp -r skills-src/skills/nlpm/orchestration .claude/skills/orchestration && rm -rf skills-src

Use ~/.claude/skills/ instead of .claude/skills for a personal install. The folder must contain SKILL.md.

Claude Code skills documentation · loads skills from .claude/skills/

Facts

Skill name
orchestration
GitHub stars
146
Token cost
~3k tokens
SKILL.md length
872 words
Files
1
Skills in repo
15
Repo updated
First seen
Licence
ISC

At a glance

Multi-agent workflow patterns: parallel dispatch, pipelines, QC gates, retries.

  • Works in 7 steps: Four Orchestration Patterns → Shared Partials for DRY → Cost Gates → …
  • Tasks that involve Multi-agent orchestration
  • SKILL.md covers 1. Four Orchestration Patterns, 2. Shared Partials for DRY, 3. Cost Gates and 4. Pipeline State, plus 3 more sections
  • Instructions only: no scripts, shell commands, URLs or credentials in SKILL.md

What it does

Orchestration is an agent skill from xiaolai/nlpm. Multi-agent workflow patterns: parallel dispatch, pipelines, QC gates, retries.

Its SKILL.md is about 3k tokens, which your agent loads only when the skill is triggered. It is a single SKILL.md file with no bundled scripts.

It sits in Agent Workflows, covering Multi-agent orchestration. The repository describes itself as: Natural-Language Programming Manager — scan, lint, and score NL artifacts with Claude-native quality scoring. The licence is ISC.

When your agent uses it

  • Tasks that involve Multi-agent orchestration

Example prompts

  • “/orchestration”

Workflow steps

7 steps, taken from the step headings in SKILL.md.

  1. Four Orchestration Patterns
  2. Shared Partials for DRY
  3. Cost Gates
  4. Pipeline State
  5. Model Tier Allocation
  6. Error Propagation
  7. Pattern Selection Guide

What it can do on your machine

Read from SKILL.md and the folder at commit 6fdbd05. It shows what the files ask for, not the result of running them.

  • Tool permissions

    Pre-approves nothing: there is no allowed-tools line, so your agent's usual permission prompts apply.

    From allowed-tools in the SKILL.md frontmatter.

  • Runs code

    No scripts in the folder and no shell commands in SKILL.md (its code samples are markdown, yaml and json).

    From the folder's file list and the shell code blocks in SKILL.md.

  • Network

    No URLs in SKILL.md.

    From URLs in SKILL.md, links to its own repository left out.

  • Credentials

    Names no API keys, tokens, secrets or passwords.

    From names ending in _API_KEY, _TOKEN, _SECRET, _KEY or _PASSWORD in SKILL.md.

Context cost

Orchestration loads about 3k tokens when it runs. Until then it costs about 23 tokens; SKILL.md has 872 words of instructions outside code blocks.

Always · name and description, kept in context so the agent knows when to use it
~23
When it runs · the whole SKILL.md, loaded when a task matches
~3k

Estimates: characters ÷ 4, the usual rule of thumb; real counts depend on the model's tokenizer. Scripts and assets cost tokens only if the agent reads them.

Safety

Auto-check passed

The automated check found no risky patterns in SKILL.md.

Automated static check — not a guarantee. Review scripts before installing. It scans the text of SKILL.md for risky patterns (piping downloads into a shell, reading credential files, hidden Unicode, destructive commands); files beside SKILL.md are not scanned.

SKILL.md

The full file from xiaolai/nlpm at commit 6fdbd05, republished under its ISC licence (© xiaolai). 872 words, ~2,955 tokens.

Download SKILL.mdSave it as .claude/skills/orchestration/SKILL.md (or your agent's skills folder).
name
orchestration
description
Multi-agent workflow patterns: parallel dispatch, pipelines, QC gates, retries.
version
0.1.0
user-invocable
false

Orchestration

Scope: covers multi-agent workflow design. For individual agent authoring, see [[writing-agents]]. For plugin architecture, see [[writing-plugins]].

1. Four Orchestration Patterns

Pattern A: Parallel Dispatch

Multiple agents run simultaneously on independent work. A command dispatches them via the Task tool and synthesizes results.

Command dispatches via Task:
  |-- agent-1 (analyzes security)
  |-- agent-2 (analyzes performance)
  |-- agent-3 (analyzes architecture)
  --> Command synthesizes all results into final report

Use when: agents don't depend on each other's output.

Real examples:

  • grill plugin: 6 review agents analyze code from different angles in parallel
  • docs-guardian: 4 agents (staleness, accuracy, coverage, quality) run simultaneously

Implementation pattern in command body:

markdown
## Execution
1. Dispatch the following agents in parallel using Task:
   - security-agent: analyze for vulnerabilities
   - performance-agent: analyze for bottlenecks
   - architecture-agent: analyze for structural issues
2. Collect all agent outputs
3. Synthesize into a unified report with cross-references

Key decisions:

DecisionRecommendation
Max parallel agents6 (diminishing returns above this)
Timeout per agent120 seconds for sonnet, 300 for opus
Failure handlingContinue with other agents if one fails
Result mergingDeduplicate findings that appear in multiple agents
Pattern B: Sequential Pipeline

Each stage feeds into the next. Output of stage N is input to stage N+1.

parse --> chunk --> summarize --> QC --> output

Use when: each stage depends on the previous stage's output.

Real examples:

  • reading-assistant: parse PDF -> chunk content -> summarize chunks -> QC summaries -> output
  • tdd-guardian: discover tests -> run tests -> check coverage -> analyze failures -> report -> enforce

Implementation pattern:

markdown
## Execution
### Phase 1: Parse (haiku)
1. Scan input files
2. Extract structured content
3. Output: parsed data as JSON

### Phase 2: Process (sonnet)
4. Receive parsed data from Phase 1
5. Analyze and transform
6. Output: processed results

### Phase 3: QC (sonnet)
7. Verify Phase 2 output meets quality bar
8. Output: pass/warn/fail verdict

### Phase 4: Output
9. If QC passed: format and deliver final report
10. If QC failed: report failures and stop

Key decisions:

DecisionRecommendation
Phase boundaryEach phase should have a clear input type and output type
Error propagationFail fast -- don't continue past a failed phase
State passingUse structured output (JSON) between phases
ResumabilityTrack phase status for long pipelines (see section 4)
Pattern C: QC Gate

AI processing followed by quality verification before output reaches the user.

Phase 1: Mechanical prep (haiku)
Phase 2: AI work (sonnet)
Phase 3: QC verification (sonnet/opus)
  --> pass: proceed to output
  --> warn: output with warnings
  --> fail: stop, report issues
Phase 4: Output

Use when: AI output needs verification before the user sees it.

Threshold design:

VerdictConditionAction
PASSAll checks greenDeliver output directly
WARNMinor issues (< 3 low-severity)Deliver output with warnings section
FAILAny critical finding OR > 5 total findingsStop, report what failed, suggest re-run

Real examples:

  • reading-assistant: QC agents verify summaries for accuracy, completeness, fidelity
  • codex-toolkit audit-fix: audit -> fix -> verify loop
Pattern D: Retry Loop

On failure, re-dispatch with error context. The agent gets a second chance with specific feedback about what went wrong.

agent produces output
  --> QC checks output
    --> pass: done
    --> fail: re-dispatch agent with error context
      --> QC re-checks
        --> pass: done
        --> fail (attempt 2): re-dispatch again
          --> max retries reached: fail with report

Use when: quality failures are recoverable by re-trying with more context.

Implementation:

markdown
## Retry Protocol
- Max retries: 3
- On retry, include in the agent prompt:
  - Previous output (or summary if too long)
  - Specific failures from QC
  - Instruction: "Fix ONLY the listed failures. Do not change passing sections."
- If max retries exhausted: output best attempt with failure annotations

Key decisions:

DecisionRecommendation
Max retries3 (rarely succeeds after 3 if it failed 3 times)
Error contextInclude specific failures, not "try again"
Scope of retryFix only failures, preserve passing output
Cost capEach retry costs full agent invocation -- budget accordingly

2. Shared Partials for DRY

Extract common logic into commands/shared/*.md with user-invocable: false in frontmatter.

When to Extract
SituationExtract?
Same logic in 3+ commandsYes -- always extract
Same logic in 2 commands, complex (> 20 lines)Yes -- extract
Same logic in 2 commands, simple (< 10 lines)No -- duplication is fine
Logic used by 1 command but might be reusedNo -- wait until it's actually reused
Good Candidates for Extraction
PartialWhat it containsWho includes it
shared/load-config.mdRead and validate plugin config fileAll commands that need config
shared/discover-files.mdFind target files by pattern/extensionCommands that scan the repo
shared/validate-prereqs.mdCheck tool availability, environmentCommands with external dependencies
shared/format-report.mdCommon report header, footer, severity colorsCommands that output reports
Partial File Structure
yaml
---
user-invocable: false
description: "Shared config loading logic — reads and validates the plugin config file"
---
markdown
## Config Loading

1. Look for `.config.md` in the project root
2. If not found, look for `.config.yaml`
3. If neither found, output error: "Run `/plugin:init` first to create a config file"
4. Parse the config file
5. Validate required fields: [list fields]
6. Return parsed config

3. Cost Gates

For expensive AI pipelines, add a cost estimation step between mechanical prep and AI processing.

Implementation
Phase 1: Parse and discover (haiku -- cheap)
  --> Count items to process
  --> Estimate cost: items x model cost per item
  --> Display estimate to user

User confirms or adjusts scope

Phase 2: AI processing (sonnet/opus -- expensive)
  --> Process confirmed scope
Show full SKILL.md (355 more words)Show less
Cost Estimation Table
ModelApprox cost per item10 items100 items1000 items
haiku$0.001$0.01$0.10$1.00
sonnet$0.01$0.10$1.00$10.00
opus$0.03$0.30$3.00$30.00

"Item" = one agent invocation processing one unit of work (one file, one chunk, one artifact).

User Confirmation Pattern
markdown
## Cost Gate
After Phase 1, display:
- Items to process: {N}
- Estimated model: {model}
- Estimated cost: ~${amount}
- Estimated time: ~{minutes} minutes

Ask: "Proceed with {N} items? (You can reduce scope with --filter)"

4. Pipeline State

For resumable pipelines (long-running, expensive, or failure-prone), track state in a JSON file.

State File Schema
json
{
  "pipeline": "my-pipeline",
  "startedAt": "2024-01-15T10:00:00Z",
  "configFingerprint": "sha256:abc123",
  "phases": {
    "parse": {
      "status": "completed",
      "startedAt": "2024-01-15T10:00:00Z",
      "completedAt": "2024-01-15T10:00:05Z",
      "itemsProcessed": 42,
      "output": "parse-output.json"
    },
    "analyze": {
      "status": "running",
      "startedAt": "2024-01-15T10:00:06Z",
      "itemsProcessed": 15,
      "itemsTotal": 42
    },
    "qc": {
      "status": "pending"
    }
  },
  "lock": {
    "pid": 12345,
    "acquiredAt": "2024-01-15T10:00:00Z"
  }
}
State Transitions
pending --> running --> completed
                   --> failed
                   --> skipped (if previous phase failed)
Resumability Rules
  1. On start: check for existing state file
  2. If state exists and configFingerprint matches: resume from last incomplete phase
  3. If state exists and configFingerprint differs: warn user, offer fresh start or resume
  4. If lock exists: check if PID is alive. If dead, clear stale lock. If alive, abort.

5. Model Tier Allocation

Assign models by cognitive load, not by importance.

Mechanical / IO:     haiku    (parser, scanner, formatter, counter)
Reasoning / AI:      sonnet   (summarizer, extractor, reviewer, linter)
Judgment / QC:       opus     (coordinator, architect, final reviewer)
Pipeline Model Assignment Example
Phase 1: Discover files          → haiku  (just glob + read)
Phase 2: Parse and chunk         → haiku  (mechanical splitting)
Phase 3: Analyze each chunk      → sonnet (requires judgment)
Phase 4: QC all analyses         → sonnet (verify, not create)
Phase 5: Synthesize final report → opus   (cross-reference, prioritize)
Cost Optimization
OptimizationHowSavings
Batch mechanical workOne haiku call processes all files, not one per file5-10x
Pre-filter before AIUse grep/glob to skip irrelevant files before sonnet2-5x
Cache phase outputsDon't re-run completed phases on retry1-3x
Scope reductionLet user filter to subset before expensive phasesVariable

6. Error Propagation

Rules
  1. Phase fails -> STOP. Set status "failed" + error message in state. Do not continue to next phase.
  2. Agent fails -> report and continue (in parallel dispatch). One agent's failure shouldn't block others.
  3. Retry fails -> escalate. After max retries, surface the failure to the user with full context.
  4. Never swallow errors silently. Every failure must be visible in the final output.
Error Report Format
markdown
## Pipeline Error

**Phase**: {phase_name}
**Status**: FAILED
**Error**: {error_message}

### Context
- Items processed before failure: {N} of {M}
- Last successful item: {item_id}
- Time elapsed: {duration}

### Recovery Options
1. Fix the finding and run `/command --resume` to continue from this phase
2. Run `/command --restart` to start fresh
3. Run `/command --skip-phase {phase_name}` to skip this phase (not recommended)
Fallback Paths

Always offer a manual fallback when automation fails:

markdown
## Fallback
If the pipeline fails after 3 retries:
1. Output all successfully processed items
2. List failed items with error context
3. Suggest manual analysis for failed items

7. Pattern Selection Guide

Your situationPatternWhy
Multiple independent analyses of same inputA: ParallelNo dependencies, maximize throughput
Each step needs previous step's outputB: SequentialData flows in one direction
AI output must be verified before deliveryC: QC GateCatch errors before user sees them
Quality failures are recoverable with feedbackD: RetryCheaper than manual re-run
Complex multi-stage with verificationB + CPipeline with QC gates between expensive phases
Multiple analyses with quality barA + CParallel dispatch, then QC all results

© xiaolai, ISC. Rendered from Markdown: HTML in the file is shown as text, images as links, and headings moved down two levels. Raw file

Files

Just SKILL.md in skills/nlpm/orchestration of xiaolai/nlpm.

Open the folder on GitHubat commit 6fdbd05

Compare with similar skills

Orchestration next to the 5 skills that share the most tags, products or categories with it. Stars are the repository's; “used in” counts other GitHub owners with a copy.

Orchestration compared with similar skills
SkillStarsUsed inTokensAuto-checkLicenceRepo updated
Orchestration this skillxiaolai/nlpm146—~3kAutomated safety check: PassISC
Orca CLIstablyai/orca87k2 repos~593Automated safety check: PassMIT
Paseo Advisor Second Opiniongetpaseo/paseo20k1 repos~756Automated safety check: PassCustom licence
O2 Review Loopopenobserve/openobserve22k—~3.7kAutomated safety check: PassAGPL-3.0
Paseo Committeegetpaseo/paseo20k1 repos~496Automated safety check: PassCustom licence
Mission Control Agent APIbuilderz-labs/mission-control6.3k—~2.1kAutomated safety check: PassMIT

Similar skills

  • Orca CLI

    stablyai/orca

    Operate Orca-managed worktrees, folder contexts, terminals, repos, automations, artifacts, skill sharing, worktree comments, and Orca's embedded browser…

    87k GitHub starsUsed in 2 repos~593 tokens
    Agent WorkflowsAuto-check passed

  • Paseo Advisor Second Opinion

    getpaseo/paseo

    Launches one separate agent through Paseo to give a second opinion on the current task, with a self-contained briefing and no permission to edit files.

    20k GitHub starsUsed in 1 repo~756 tokens
    Agent WorkflowsAuto-check passed

  • O2 Review Loop

    openobserve/openobserve

    Splits a change into planner, coder and independent reviewer roles: you confirm a spec, a subagent implements it, and a separate reviewer checks each round's local WIP commit.

    22k GitHub stars~3.7k tokensUpdated today
    Agent WorkflowsAuto-check passed

  • Paseo Committee

    getpaseo/paseo

    Forms a two-agent committee with contrasting profiles to analyze a stuck problem in parallel, reconcile their views and return a consensus plan without editing files.

    20k GitHub starsUsed in 1 repo~496 tokens
    Agent WorkflowsAuto-check passed

  • Mission Control Agent API

    builderz-labs/mission-control

    Teaches an agent to use the Mission Control dashboard API: register, send heartbeats, fetch assigned tasks, report progress and disconnect, with API key auth.

    6.3k GitHub stars~2.1k tokensUpdated 10 days ago
    Agent WorkflowsAuto-check passed

  • Paseo Agent Handoff

    getpaseo/paseo

    Hands off the current task, including context, decisions and failed attempts, to a fresh agent through Paseo by writing a self-contained briefing prompt and launching that agent.

    20k GitHub starsUsed in 1 repo~606 tokens
    Agent WorkflowsAuto-check passed

More from xiaolai/nlpm

All 15 skills in this repo

  • Conventions

    xiaolai/nlpm

    Universal NL conventions: SKILL.md open spec, AGENTS.md, vague quantifiers, naming.

    146 GitHub stars~3.6k tokensUpdated today
    Auto-check passed

  • Conventions Antigravity

    xiaolai/nlpm

    Antigravity and Gemini CLI artifact schemas: .gemini/ paths, extensions, hooks.

    146 GitHub stars~3.1k tokensUpdated today
    Auto-check passed

  • Conventions Codex

    xiaolai/nlpm

    Codex CLI artifact schemas: config.toml, .codex-plugin, skills, hooks, AGENTS.md.

    146 GitHub stars~4.9k tokensUpdated today
    Auto-check passed

  • Patterns

    xiaolai/nlpm

    NL artifact anti-patterns: vague quantifiers, bare prohibitions, oversized skills.

    146 GitHub stars~3.9k tokensUpdated today
    Auto-check passed

  • Scoring

    xiaolai/nlpm

    100-point NL artifact rubric: penalty tables per artifact type, calibration cases.

    146 GitHub stars~5.3k tokensUpdated today
    Auto-check passed

  • Testing

    xiaolai/nlpm

    NL artifact test specs for /nlpm:test: spec format, TDD for skills and agents.

    146 GitHub stars~1.4k tokensUpdated today
    Auto-check passed

Categories

Questions about Orchestration

What does Orchestration do?

Multi-agent workflow patterns: parallel dispatch, pipelines, QC gates, retries. Orchestration is an agent skill from xiaolai/nlpm. Multi-agent workflow patterns: parallel dispatch, pipelines, QC gates, retries.

When should I use Orchestration?

Orchestration fits situations like: tasks that involve Multi-agent orchestration.

How do I install Orchestration in Claude Code?

Run `npx skills add xiaolai/nlpm --skill orchestration -a claude-code`. Or copy the skill folder (skills/nlpm/orchestration in xiaolai/nlpm) into .claude/skills/orchestration in your project. Claude Code loads it when a task matches its description.

How do I install Orchestration in Codex?

Run `npx skills add xiaolai/nlpm --skill orchestration -a codex`. Or copy the skill folder (skills/nlpm/orchestration in xiaolai/nlpm) into .agents/skills/orchestration in your project. Codex loads it when a task matches its description.

Can I use Orchestration in Cursor, Gemini CLI or GitHub Copilot?

Cursor, Gemini CLI, GitHub Copilot and OpenCode also load SKILL.md folders. With the skills CLI, run `npx skills add xiaolai/nlpm --skill orchestration -a cursor` (or -a gemini-cli, github-copilot or opencode for the others). To copy it by hand, put the folder in .cursor/skills/orchestration, .gemini/skills/orchestration, .github/skills/orchestration and .opencode/skills/orchestration in your project.

What does Orchestration need to run?

SKILL.md names no scripts, command-line tools or credentials: Orchestration is instructions for the agent only.

Does Orchestration access the network?

SKILL.md contains no URLs. Any network use would come from the scripts or tools the agent runs. This is read from the text; nothing was executed.

Is Orchestration safe to install?

Our automated static check of SKILL.md found no risky patterns, such as piping downloads into a shell, reading credential files or hidden Unicode. It is not a guarantee. Review the folder before installing.

What licence does Orchestration use?

Orchestration is published under the ISC licence (the repository's licence). It allows redistribution, so the full SKILL.md is shown on this page.

How many tokens does Orchestration use?

About 3k tokens (SKILL.md is roughly 12k characters). Agents keep only the skill's name and description in context until a task matches; then they load SKILL.md in full.

What are the alternatives to Orchestration?

Skills that share tags, products or a category with Orchestration: Orca CLI (stablyai/orca, 87k stars), Paseo Advisor Second Opinion (getpaseo/paseo, 20k stars), O2 Review Loop (openobserve/openobserve, 22k stars) and Paseo Committee (getpaseo/paseo, 20k stars). The comparison table on this page puts their stars, adoption, token cost, safety result and licence side by side.

Who maintains Orchestration?

xiaolai (a GitHub user) maintains it in xiaolai/nlpm, which has 146 GitHub stars. The repository holds 15 skills in this directory. The repository was last updated on October 8, 2026.

Source: xiaolai/nlpm on GitHub. Facts on this page come from the repository at the commit we read; the author's words are quoted as theirs.