nw-agent-creation-workflow

Agent Creation Workflow

Overview

Create agents through 5 phases: ANALYZE -> DESIGN -> CREATE -> VALIDATE -> REFINE. Each phase has clear inputs, outputs, and quality gates. Follow "start minimal, add based on failure."

Phase 1: ANALYZE

Goal: Understand requirements and determine agent architecture.

Inputs: User requirements, use case description, existing codebase context.

Steps:

1. Identify single clear responsibility
2. Determine new agent or modification of existing
3. Check overlap with existing agents (avoid duplication)
4. Classify agent type:
5. Specialist: Single-domain expert (most common)
6. Reviewer: Validates outputs from another agent (Reflection pattern)
7. Orchestrator: Coordinates multiple agents
8. Identify required tools (start with Read, Glob, Grep -- add only what's needed)
9. Determine if Skills needed (domain knowledge > 50 lines)

Gate: Single responsibility identified. Agent type classified. No overlap.

Output: Requirements summary with agent type, tools list, skill needs.

Phase 2: DESIGN

Goal: Design agent architecture and structure.

Inputs: Requirements summary from Phase 1.

Steps:

1. Select design pattern (load design-patterns skill)
2. Define role and goal (1-2 sentences each)
3. Identify core principles that DIVERGE from Claude defaults:
4. What must this agent do differently than Claude naturally would?
5. Domain-specific methodology steps
6. Non-obvious constraints | Project-specific conventions
7. Design workflow (3-7 phases)
8. Plan Skills extraction: domain knowledge -> separate Skill | Testing/validation -> separate Skill | Keep workflow and principles in core agent
9. Design Skill Loading Strategy (required for 3+ skills):
10. Map each skill to the workflow phase where it's needed
11. Create a loading table: Phase → Skill → Trigger condition
12. Add explicit Load: skill-name directives in each workflow phase
13. Document path: ~/.claude/skills/nw-{skill-name}/SKILL.md (installed) or nWave/skills/nw-{skill-name}/SKILL.md (repo)
14. Note: skills: in frontmatter is declarative only — Claude Code does NOT auto-load skill files. The agent must use Read tool to load them, triggered by Load: directives in workflow text.
15. Draft frontmatter:

```yaml

name: {kebab-case-id} description: Use for {domain}. {When to delegate.} model: inherit tools: [{minimum tools needed}] maxTurns: 30 skills:

• nw-{skill-name}

```

Gate: Design fits ~200-300 lines (core) + Skills. Pattern selected. Frontmatter drafted.

Output: Agent architecture document (working notes, not deliverable).

Phase 3: CREATE

Goal: Write agent definition file and Skills.

Inputs: Design from Phase 2.

Steps:

1. Create agent .md file:
2. YAML frontmatter (name, description, tools, model, maxTurns, skills)
3. Role + Goal paragraph
4. Core Principles (divergences only, 3-8 items)
5. Workflow phases
6. Critical Rules (3-5, where violation causes real harm)
7. Examples (3-5 canonical cases)
8. Subagent mode instructions
9. Constraints (what agent does NOT do)
10. Create Skill files if needed: each in nWave/skills/{agent-name}/ | YAML frontmatter with name and description | Focused content, 100-250 lines each
11. Measure: wc -l. Target: under 300 lines.

1. Add Skill Loading Strategy section (required for agents with 3+ skills):

```markdown ## Skill Loading Strategy

Load on-demand by phase, not all at once:

| Phase | Load | Trigger | |-------|------|---------| | 1 Phase Name | skill-name | Always — core methodology | | 2 Phase Name | other-skill | When condition is met |

Skills path: ~/.claude/skills/nw-{skill-name}/SKILL.md (installed) or nWave/skills/nw-{skill-name}/SKILL.md (repo) ```

1. Add Load: directives at the start of each workflow phase referencing the applicable skills
2. Verify: every skill in frontmatter skills: has at least one Load: directive in the workflow text. Orphan skills (declared but never loaded) are a bug.

Gate: Agent file created. Under 300 lines. Skills created if needed. Skill Loading Strategy present for 3+ skills.

Output: Agent .md file + Skill files.

Phase 4: VALIDATE

Goal: Verify agent meets quality standards.

Steps:

1. Run 14-point validation checklist:
2. [ ] Uses official YAML frontmatter format
3. [ ] Total definition under 400 lines (domain knowledge in Skills)
4. [ ] Only specifies behaviors diverging from Claude defaults
5. [ ] No aggressive signaling language
6. [ ] 3-5 canonical examples for critical behaviors
7. [ ] Tools restricted to minimum needed
8. [ ] maxTurns set in frontmatter
9. [ ] Safety via platform features, not prose
10. [ ] Instructions phrased affirmatively
11. [ ] Consistent terminology throughout
12. [ ] Description clearly states delegation criteria
13. Check anti-patterns: no monolithic sections (>50 lines without structure) | No duplicated Claude defaults | No embedded safety frameworks | No aggressive language
14. Test with representative inputs (Layer 1 testing)

Gate: All 14 items pass. No anti-patterns.

Output: Validation report (pass/fail per item).

Phase 5: REFINE

Goal: Iteratively improve based on testing feedback.

Steps:

1. Address validation failures
2. Test with edge cases
3. Add instructions ONLY for observed failure modes: wrong decision -> add rule/example | Missed step -> clarify workflow | Over-generated -> add constraint
4. Re-measure: wc -l. If approaching 400 lines, extract to Skills.
5. Re-validate with 14-point checklist.

Gate: All validation passes. Line count within target. Edge cases handled.

Output: Final agent definition, ready for installation.

Quality Gates Summary

| Phase | Gate | Blocks | |-------|------|--------| | ANALYZE | Single responsibility, no overlap | DESIGN | | DESIGN | Architecture fits size target | CREATE | | CREATE | File created, under 300 lines | VALIDATE | | VALIDATE | 14-point checklist passes | REFINE/Deploy | | REFINE | Edge cases handled, within target | Deploy |

Naming Conventions

• Agent files: nw-{name}.md in nWave/agents/
• Skill files: {skill-name}.md in nWave/skills/{agent-name}/
• Reviewer agents: nw-{name}-reviewer.md
• Agent names in frontmatter: nw-{name} (kebab-case with nw- prefix)

Reviewer Agent Creation (Special Case)

Reviewer agents pair with a primary agent and use the Reflection pattern:

1. Set model: haiku in frontmatter (cost-efficient review)
2. Use same tools as primary agent (no Write/Edit -- reviewers don't modify)
3. Define structured critique output format (YAML)
4. Include max 2 review iterations
5. Define clear approval/rejection criteria