t1k:agent-creator

Create or update TheOneKit agent .md files with canonical structure. Use when adding a new agent, updating maxTurns/model, or fixing frontmatter fields.

SKILL.md

GitHub

Agent Creator

Create new agent .md files or update existing ones for TheOneKit. Enforces the canonical agent structure.

Required Frontmatter

Field Required Notes
name Yes kebab-case, must match filename
description Yes 2-5 <example> blocks with <commentary>
model Yes inherit, sonnet, opus, or haiku
maxTurns Yes 15-50 based on role
color Suggested Consistent per role
origin/module/protected NO CI-injected only — never add manually

Recommended maxTurns

Role Range
implementer / kit-developer 40-50
debugger 35-40
tester 30-35
planner / brainstormer 25-30
reviewer 25-30
docs / git / utility 15-25

Required Body Sections (in order)

  1. Opening: "You are a [specific role]..."
  2. Routing guard (if applicable): scope boundary + delegation targets
  3. Mandatory skills: Ordered table with activation triggers
  4. Constraints: Enforced rules (NEVER/ALWAYS/DO NOT)
  5. Workflow: Numbered steps with verification checkpoints
  6. Output format: Structured markdown template
  7. Completion gates: Verifiable, tool-referenced, severity-marked (BLOCKING/MANDATORY)
  8. Module awareness (if applicable)

Process

  1. Ask user for: role name, scope, which kit/layer, key skills
  2. Read 2 existing agents as reference (matching role type)
  3. Draft agent following canonical structure
  4. Validate:
    • All required frontmatter present
    • 2+ examples in description
    • All 7 body sections present
    • Completion gates are verifiable
    • maxTurns justified
  5. If routing needed: update t1k-routing-{layer}.json
  6. If activation needed: update t1k-activation-{layer}.json

Cognitive Framing (MANDATORY for all new agents)

Every new agent MUST include a persona line as the first sentence of the body (after frontmatter).

Format: "You are a {Role Title} who/performing {behavioral description in 1-2 sentences}."

Examples:

  • "You are a Staff Engineer hunting production bugs — you distrust obvious answers and prove root cause before proposing any fix."
  • "You are a Detective performing systematic investigation. You form hypotheses, gather evidence, and never assume."
  • "You are a Tech Lead designing system architecture — you optimize for long-term maintainability over short-term convenience."
  • "You are a QA Lead performing systematic verification — you hunt for untested code paths and think like someone burned by production incidents."

Why it matters: The persona shapes reasoning quality throughout the agent's entire run. Generic descriptions ("you are a developer") produce generic, shallow output. Specific behavioral frames ("you distrust obvious answers") activate different reasoning patterns.

Validation: If the opening line does not name a concrete role with a behavioral trait, reject and rewrite.

Anti-Patterns to Avoid

  • Vague role ("a developer") — must be specific with behavioral trait
  • Missing cognitive framing persona as first body line
  • "Activate skills as needed" (must be explicit)
  • No completion gates
  • No routing guard when scope overlaps with other agents
  • maxTurns >50
  • Adding origin/module/protected manually