skills.rest
dvy1987
dvy1987COMMUNITY
Explore all skills by this author →
WebsiteGitHub

agent-creator

Launch subagents from validated architecture specs using Claude Code or Ampcode Task.

3|Updated Apr 5, 2026
Search Tags:#monitoring#orchestration#claude-code#task-tool#agent-launch#ampcode#launch-manifest
One-click install
npx skills add https://github.com/dvy1987/agent-loom --skill agent-creator-dvy1987
Or copy as Structured Prompt for Agent▼
Please help me install this Agent Skill.
Skill: agent-creator
Source: https://github.com/dvy1987/agent-loom/tree/main/.agents/skills/agent-creator
Command: npx skills add https://github.com/dvy1987/agent-loom --skill agent-creator-dvy1987

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

This skill automates the secure, auditable launching of subagents from a validated architecture specification, removing manual orchestration, reducing human error, and ensuring executions follow a recorded manifest and failure-handling rules.

Core Features & Use Cases

  • Precondition enforcement: verifies presence of architecture spec, a setup-evaluation PASS, and per-agent prompt files before launching.
  • Platform-aware spawn instructions: emits structured, non-code spawn manifests for Claude Code or Ampcode Task tool topologies (parallel, sequential, hierarchical).
  • Monitoring and handoff: writes launch manifests to docs/agents/runs/, polls docs/handoffs/ for outputs, logs failures with HANDOFF-FAILED files, and hands off results to project-orchestrator.

Quick Start

Launch agents for a validated architecture by running the agent-creator after confirming the architecture spec, setup-evaluation PASS, and all docs/agents/*-prompt.md files are present.

Frequently Asked Questions about agent-creator

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I launch subagents from a validated architecture specification?▼

To launch subagents, this skill verifies the presence of an architecture spec, a setup-evaluation PASS, and per-agent prompt files, then emits structured spawn manifests for Claude Code or Ampcode Task tool topologies.

What preconditions are required for agent orchestration using Claude Code?▼

Agent orchestration requires docs/architecture/*-arch.md, a recorded setup-evaluation PASS, per-agent prompt files in docs/agents, and a platform with native Task tool support to proceed.

How does hierarchical subagent orchestration handle failures and handoffs?▼

Hierarchical subagent orchestration handles failures by logging them with HANDOFF-FAILED files, polling docs/handoffs/ for outputs, and handing off results to project-orchestrator.

Can I use Ampcode Task tool for parallel subagent topologies without manual orchestration?▼

Yes, you can use Ampcode Task tool for parallel subagent topologies. The skill automates secure spawning by writing launch manifests to docs/agents/runs/ and removing manual orchestration.

Why does my agent launch fail despite having per-agent prompt files?▼

Agent launches fail if preconditions are unmet: you need a validated architecture spec in docs/architecture/, a setup-evaluation PASS, prompt files in docs/agents, and native Task tool support.

What is the best way to monitor sequential agent executions after launch?▼

The best way to monitor sequential agent executions is by checking the recorded launch manifests written to docs/agents/runs/ and polling docs/handoffs/ for execution outputs.

Related Skills

Explore More Software Engineering Skills→
ichabodcoleichabodcole

review-docs

Assign docs-curator agents to validate documentation against the codebase.

Community
Intermediate
tsilvatsilva

investigate-root-cause

Orchestrate four-agent ensemble investigations to determine health condition root causes.

Community
Advanced
proffesor-for-testingproffesor-for-testing
43

swarm-advanced

Orchestrate advanced swarm patterns for multi-agent workflows across distributed systems.

Community
Advanced
Imad-OuteImad-Oute
2

mh-cto-sdlc-go-live

Automate production launches by validating preconditions, running smoke tests, and monitoring deployment.

Community
Advanced