agentleFS
varunk130varunk130/ai-workflow-playbooks/skills/multi-agent-collaboration/SKILL.md

multi-agent-collaboration

Skill2 starsChanged 6 months agoMITView on GitHub

Coordinate effectively when multiple AI agents work on the same codebase or feature.

6 other projects use a version of this.

Adopted by

Most recently taken up by

The file

Show its rules and text

We split this skill into the rules it gives an agent

15 rules. 0 are common, 15 are distinctive, and 0 are very specific to this project.

  • 0

    Common

    Most of the 6 edited versions we read keep these.

  • 15

    Distinctive

    This author's own choices. Worth a look.

  • 0

    Very specific

    Tied to their project's names, files or tools. Probably skip.

  • Common
  • Distinctive
  • Very specific
  1. Distinctive:

    When agents produce code that connects, define the contract first:

    line 21
  2. Distinctive:

    Write the interface or type definition first, implement second

    line 24
  3. Distinctive:

    Read the project's style guide or infer it from existing code before writing

    line 29
  4. Distinctive:

    Use the same naming patterns, file structure, and error handling approach

    line 30
  5. Distinctive:

    Run the same linter and formatter configuration

    line 31
  6. Distinctive:

    Follow the same commit message format

    line 32
  7. Distinctive:

    Avoid merge conflicts proactively:

    line 35
  8. Distinctive:

    Check which files other agents have modified before starting

    line 36
  9. Distinctive:

    Prefer additive changes (new files) over modifications to shared files

    line 37
  10. Distinctive:

    When shared files must change, coordinate the order - one agent goes first

    line 38
  11. Distinctive:

    Use feature flags to integrate work independently

    line 39
  12. Distinctive:

    When passing work to another agent:

    line 42
  13. Distinctive:

    Own your scope: - Never modify files outside your assigned area without coordination

    line 55
  14. Distinctive:

    Contracts before code: - When work connects, agree on the interface first

    line 56
  15. Distinctive:

    Verify integration: - After merging parallel work, run the full test suite

    line 59

“Common” is measured: kept by at least 6 in 10 of the 6 edited versions of this file we could read. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.

Keep what fits and drop the rest. Your version, ready for your agents.
Files
SKILL.md

Multi-Agent Collaboration

What This Skill Enables

An agent that can operate alongside other agents - dividing work, avoiding conflicts, and maintaining consistency - without stepping on each other's changes or duplicating effort. Without this skill, multi-agent workflows produce merge conflicts, inconsistent patterns, and wasted cycles.

Core Competencies

1. Work Partitioning

Divide tasks so agents work on independent surfaces:

  • By file: Agent A owns auth/, Agent B owns dashboard/
  • By layer: Agent A handles API endpoints, Agent B handles UI components
  • By feature slice: Agent A builds user creation end-to-end, Agent B builds search end-to-end
  • Never: Two agents modifying the same file simultaneously

2. Contract Agreement

When agents produce code that connects, define the contract first:

  • Agree on interface signatures (function names, parameter types, return types)
  • Agree on data shapes (request/response schemas, event payloads)
  • Write the interface or type definition first, implement second
  • Both agents can work in parallel once the contract is locked

3. Convention Adherence

Multiple agents must produce code that looks like one person wrote it:

  • Read the project's style guide or infer it from existing code before writing
  • Use the same naming patterns, file structure, and error handling approach
  • Run the same linter and formatter configuration
  • Follow the same commit message format

4. Conflict Prevention

Avoid merge conflicts proactively:

  • Check which files other agents have modified before starting
  • Prefer additive changes (new files) over modifications to shared files
  • When shared files must change, coordinate the order - one agent goes first
  • Use feature flags to integrate work independently

5. Handoff Protocol

When passing work to another agent:

HANDOFF NOTE:
- What I completed: [list of files/features implemented]
- What I tested: [which tests pass, what was verified]
- What remains: [tasks not yet started]
- Assumptions I made: [decisions the next agent should know about]
- Known issues: [anything broken or incomplete]
- Files to read first: [most important context files]

Behavioral Rules

  1. Own your scope - Never modify files outside your assigned area without coordination
  2. Contracts before code - When work connects, agree on the interface first
  3. Commit frequently - Small, frequent commits reduce merge conflict surface area
  4. Document decisions - Other agents can't read your reasoning; write it down
  5. Verify integration - After merging parallel work, run the full test suite

Failure Modes

FailureSymptomCorrection
Both agents edit the same fileMerge conflicts, lost changesPartition by file or coordinate editing order
Inconsistent patternsCode looks like two different projectsAlign on conventions before starting
Missing handoff contextNext agent repeats work or breaks assumptionsUse the handoff protocol template
No integration testingIndividual pieces work, combined system doesn'tRun full suite after merging any parallel work

Download now (ZIP)

Does it work?

Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.

No reviews yet

People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.

  • It worked
  • Partly
  • It didn't
What did you use it for? e.g. Reviewing pull requests in a Rails app

Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.

Your trail

Files you open will collect here, so you can find them again.

Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.

Now do this for your own.

Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.