agentleFS
d-o-hubd-o-hub/github-template-ai-agents/.agents/skills/goap-agent/SKILL.md

goap-agent

Skill2 starsChanged 11 days agoMITView on GitHub

Orchestrates complex multi-step tasks with intelligent planning: analyze the problem, decompose into sub-goals, select execution strategy, assign agents, and coordinate with quality gates. Use this skill when the user asks to plan a large change, break down a complex problem, coordinate multiple agents, or systematically tackle a multi-file refactoring — even if they just say "plan this out" or "how should we approach this". Not for simple single-step tasks (use delegate) or implementing from an approved plan (use implementer).

1 other project uses a version of this, 1 of them edited.

Adopted by

Most recently taken up by

The file

We split this skill into the rules it gives an agent

5 rules. About 0 are common, 3 are distinctive, and 2 are very specific to this project.

  • ≈0

    Common · estimate

    Many files like this have a rule of this kind.

  • ≈3

    Distinctive · estimate

    This author's own choices. Worth a look.

  • 2

    Very specific

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

  • Common
  • Distinctive
  • Very specific
  1. Very specific:

    Always use the plans/ folder for all files.

    line 19
  2. Distinctive:

    Use this skill when facing:

    line 28
  3. Distinctive:

    Before decomposing tasks, ensure architectural decisions are sound:

    line 44
  4. Very specific:

    Analyze: Use triz-solver (solve or audit mode) to evaluate the problem and resolve contradictions.

    line 45
  5. Distinctive:

    Gate: Wait for human approval of the ADR before proceeding to decomposition.

    line 47

≈ “Common” is an estimate: at least 1 in 5 of the 350,495 skill files we read have a rule of the same kind, not the same words. “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

GOAP Agent Skill: Goal-Oriented Action Planning

Enable intelligent planning and execution of complex multi-step tasks through systematic decomposition, dependency mapping, and coordinated multi-agent execution.

Always use the plans/ folder for all files. Use plans/GOAP_STATE.md to track persistent state.

Quick Reference

  • execution-strategies.md - Detailed guide on execution patterns
  • references/guide.md - Complete examples, templates, and advanced topics

When to Use

Use this skill when facing:

  • Complex Multi-Step Tasks: Tasks requiring 5+ distinct steps
  • Cross-Domain Problems: Issues spanning multiple areas
  • Optimization Opportunities: Tasks benefiting from parallel execution
  • Quality-Critical Work: Projects requiring validation checkpoints

Core GOAP Methodology

The GOAP Planning Cycle

1. ANALYZE → 2. DECOMPOSE → 3. STRATEGIZE → 4. COORDINATE → 5. EXECUTE → 6. SYNTHESIZE

Phase 1: Task Analysis (Analyze & Strategize)

Before decomposing tasks, ensure architectural decisions are sound:

  1. Analyze: Use triz-solver (solve or audit mode) to evaluate the problem and resolve contradictions.
  2. Decide: Formulate an ADR (Architecture Decision Record) detailing Context, Decision, and Consequences.
  3. Gate: Wait for human approval of the ADR before proceeding to decomposition.
## Task Analysis

**Primary Goal**: [Clear statement of what success looks like]
**Constraints**: [Time, Resources]
**Complexity**: Simple/Medium/Complex
**ADR Link**: [Link to approved Architecture Decision Record]

Context: Use Explore agent, check past patterns, perform TRIZ analysis, and record architectural decisions.

Phase 2: Task Decomposition

Decompose high-level objectives into manageable, testable sub-tasks.

Decomposition Framework

  1. Requirements Analysis: Extract primary objective, implicit requirements, constraints, success criteria.
  2. Goal Hierarchy: Top-down decomposition into sub-goals and atomic tasks.
  3. Dependency Mapping: Sequential (A→B→C), Parallel (A,B,C), Converging (A,B,C→D).
  4. Success Criteria: Define inputs, outputs, quality standards.

Atomic Criteria

Each task must be: Single action, defined inputs/outputs, one agent, testable.

Decomposition Patterns

  • Layer-Based: Data → Business Logic → API → Testing → Docs
  • Feature-Based: Core MVP → Error Handling → Performance → Integration → Testing → Docs
  • Phase-Based: Research → Foundation → Implementation → Integration → Polish → Release
  • Problem-Solution: Reproduce → Diagnose → Design → Fix → Verify → Prevent

Output Format

### Sub-Goals
1. [Component 1] - Priority: P0, Deps: none
2. [Component 2] - Priority: P1, Deps: Component 1

Principles: Atomic, Testable, Independent, Assigned. No task >4 hours.

Phase 3: Strategy Selection

StrategyWhenSpeed
ParallelIndependent tasksNx
SequentialDependent tasks1x
SwarmMany similar tasks~Nx
HybridMixed requirements2-4x

See execution-strategies.md for details.

Phase 4: Agent Assignment

AgentBest For
feature-implementerNew functionality
debuggerBug fixes
test-runnerTest validation
refactorerCode improvements
code-reviewerQuality assurance
delegateContext retrieval, pattern search, state assessment (before planning)
implementerAtomic code changes post-Blueprint approval

Phase 5: Execution Planning

## Execution Plan

- Strategy: [Type]
- Quality Gates: [N checkpoints]

### Phase 1
- Tasks: [List]
- Quality Gate: [Criteria]

Phase 6: Coordinated Execution

Parallel: Single message, multiple Task tool calls Sequential: Phases with quality gates between Monitor: Track progress, validate results

Phase 7: Result Synthesis

## Summary

✓ Completed: [Tasks]
📦 Deliverables: [List]
✅ Quality: [Status]

Common Patterns

  • Research → Implement → Validate
  • Investigate → Diagnose → Fix → Verify
  • Audit → Improve → Validate

Error Handling

  • Agent Failure: Retry, Reassign, Modify, or Escalate
  • Quality Gate Failure: Re-run with fixes
  • Blocked: Re-order or work on independent tasks

Best Practices

DO:

✓ Break tasks into atomic units ✓ Define clear quality gates ✓ Match agents to requirements ✓ Monitor and validate incrementally

DON'T:

✗ Create monolithic tasks ✗ Skip quality gates ✗ Assume independence without verification

Integration

  • execution-strategies.md: Strategy implementation (parallel, sequential, swarm)

Summary

GOAP enables systematic planning through: Analysis, Decomposition, Strategy, Quality Assurance, and Coordinated Agents.

See Also

  • triz-solver — Solve problems and audit systems for contradictions using TRIZ

Rationalizations

RationalizationReality
"I can just execute all tasks in parallel — it's faster."Parallel execution without dependency analysis causes race conditions, wasted tokens, and corrupted intermediate state.
"Quality gates slow things down — skip them for speed."Skipping quality gates compounds errors across agents, requiring far more rework than the time saved.

Red Flags

  • Executing tasks without mapping dependencies first
  • Skipping ADR approval gate before decomposition
  • Assigning agents without matching skills to task requirements

References

  • references/guide.md - Complete templates, detailed examples, extended patterns, error handling, optimization
  • execution-strategies.md - Execution pattern details

Voice & Context

  • Default: professional + blog
  • Reference: voice-profiles skill for definitions and auto-detection.

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.