Skill

create-sub-agent

Design and scaffold a Claude Code sub-agent

Install

Run in your terminal

npx claudepluginhub richfrem/agent-plugins-skills --plugin agent-scaffolders

Tool Access

This skill is limited to using the following tools:

BashReadWrite

Supporting Assets

View in Repository

acceptance-criteria.md

evals/evals.json

evals/results.tsv

fallback-tree.md

references/acceptance-criteria.md

references/examples/agent-creation-prompt.md

references/examples/complete-agent-examples.md

references/examples/validate-agent.sh

references/fallback-tree.md

references/hitl-interaction-design.md

references/pattern-decision-matrix.md

references/patterns/action-forcing-output-with-deadline-attribution.md

references/patterns/adversarial-objectivity-constraint.md

references/patterns/anti-pattern-vaccination.md

references/patterns/anti-symptom-triage.md

references/patterns/artifact-embedded-execution-audit-trail.md

references/patterns/artifact-generation-xss-compliance-gate.md

references/patterns/artifact-lifecycle.md

references/patterns/artifact-state-interrogative-routing.md

references/patterns/asynchronous-benchmark-metric-capture.md

Skill Content

Similar Skills

executing-plans

Executes pre-written implementation plans: critically reviews, follows bite-sized steps exactly, runs verifications, tracks progress with checkpoints, uses git worktrees, stops on blockers.

superpowers

134.2k

brainstorming

7 files

Guides idea refinement into designs: explores context, asks questions one-by-one, proposes approaches, presents sections for approval, writes/review specs before coding.

superpowers

134.2k

dispatching-parallel-agents

Dispatches parallel agents to independently tackle 2+ tasks like separate test failures or subsystems without shared state or dependencies.

superpowers

134.2k

Stats

Parent Repo Stars1

Parent Repo Forks1

Last CommitApr 5, 2026

Actions

View Source View Plugin View on GitHub View README

Steps

If $ARGUMENTS is provided, use it as the starting context for agent name / purpose

Follow the create-sub-agent phased workflow: extract core intent via design interview (purpose, input/output contract, escalation posture, tools, permissions.deny, model, maxTokens, color, lifecycle hooks, placement), present design summary, confirm, then generate the agent .md file

Validate the generated agent with validate_agent.py

Report the created agent path, triggering conditions, and next steps

Edge Cases

If $ARGUMENTS is empty: conduct the full Phase 1 design interview — do not pre-fill

If an agent with that name already exists: confirm before overwriting

If requirements suggest multiple responsibilities: propose splitting into specialized agents

If high-risk operations are required: configure escalation posture and add Stop hook

Follow the create-sub-agent skill workflow to design and generate a Claude Code agent file.

Inputs

$ARGUMENTS — optional agent name or brief use-case description passed as initial context to the design interview. Omit to start with open discovery.

Steps

If $ARGUMENTS is provided, use it as the starting context for agent name / purpose
Follow the create-sub-agent phased workflow: extract core intent via design interview (purpose, input/output contract, escalation posture, tools, permissions.deny, model, maxTokens, color, lifecycle hooks, placement), present design summary, confirm, then generate the agent .md file
Validate the generated agent with validate_agent.py
Report the created agent path, triggering conditions, and next steps

Output

Agent .md file with complete YAML frontmatter (name, description with <example> blocks, model, maxTokens, color, permissions.allowedTools, permissions.deny) and a second-person system prompt targeting 500-3,000 characters.

Edge Cases

If $ARGUMENTS is empty: conduct the full Phase 1 design interview — do not pre-fill
If an agent with that name already exists: confirm before overwriting
If requirements suggest multiple responsibilities: propose splitting into specialized agents
If high-risk operations are required: configure escalation posture and add Stop hook