About

Imported from kjuhwa/skills-hub (skills/workflow/split-llm-codegen-into-idea-plus-per-item/SKILL.md). Install upstream with npx skills add kjuhwa/skills-hub --skill split-llm-codegen-into-idea-plus-per-item. Copyright stays with the author.

Split LLM Code Generation: Idea + Per-Item Calls

Name: split-llm-codegen-into-idea-plus-per-item
Author: kjuhwa

When generating multiple large code artifacts via LLM CLI, a single call requesting N items produces N small results. Splitting into (1 + N) calls — one lightweight idea call plus N individual code calls — gives each artifact the full response capacity.

When to use

Generating 3+ substantial code artifacts (apps, components, modules) in one workflow
Each artifact needs 500+ lines
Single-call approach produces undersized results

Pattern

Call 1 (lightweight, 2min timeout):
  "Generate N ideas with names, descriptions, features"
  → Parse N idea blocks

Call 2..N+1 (heavy, 30min timeout each):
  "Build ONE complete app for idea[i] with full spec"
  → Each gets full response capacity

Key decisions

Idea call is cheap (short response) — use tight timeout (60-120s)
Code calls are heavy — use generous timeout (15-30min)
Pass idea metadata (name, features) into each code call for consistency
Concatenate results after all calls complete

Anti-patterns

Requesting all artifacts in one call ("generate 3 complete 2000-line apps") — each gets ~1/3 capacity
Not parsing idea output before code calls — fails silently if ideas malformed
Equal timeouts for idea vs code calls — wastes time or causes premature timeout

Use it

Copy one of these into your project. Installing also returns the manifest and these snippets.

yaml

targets:
  - https://api.opensmartroute.ai/api/v1/registry/kjuhwa-skills-hub-split-llm-codegen-into-idea-plus-per-item/manifest   # or paste the manifest below

Manifest

An Open Capability Manifest: the router reads it to know what this does, what it costs and when to pick it.

kjuhwa-skills-hub-split-llm-codegen-into-idea-plus-per-item.ocm.jsonjson

{
  "ocm": "1",
  "id": "kjuhwa-skills-hub-split-llm-codegen-into-idea-plus-per-item",
  "kind": "skill",
  "name": "split-llm-codegen-into-idea-plus-per-item",
  "description": "Split large LLM code generation into lightweight idea call + per-item full code calls to maximize output per artifact",
  "publisher": "kjuhwa",
  "version": "1.0.0",
  "capabilities": {
    "domains": [
      "coding"
    ],
    "tags": [
      "skill-md",
      "llm",
      "code-generation",
      "output-limit",
      "multi-call",
      "github"
    ],
    "languages": [
      "en"
    ]
  },
  "quality_prior": 0.6,
  "examples": [
    "Split large LLM code generation into lightweight idea call + per-item full code calls to maximize output per artifact"
  ],
  "primary": false,
  "metadata": {
    "source": {
      "provider": "github",
      "repository": "https://github.com/kjuhwa/skills-hub",
      "path": "skills/workflow/split-llm-codegen-into-idea-plus-per-item/SKILL.md",
      "ref": "b8e7275ea024cc3f9d8551b41aa9ebf89b464cd1",
      "url": "https://github.com/kjuhwa/skills-hub/blob/b8e7275ea024cc3f9d8551b41aa9ebf89b464cd1/skills/workflow/split-llm-codegen-into-idea-plus-per-item/SKILL.md",
      "key": "kjuhwa/skills-hub/skills/workflow/split-llm-codegen-into-idea-plus-per-item/SKILL.md"
    }
  },
  "instructions": "# Split LLM Code Generation: Idea + Per-Item Calls\n\nWhen generating multiple large code artifacts via LLM CLI, a single call requesting N items produces N small results. Splitting into (1 + N) calls — one lightweight idea call plus N individual code calls — gives each artifact the full response capacity.\n\n## When to use\n\n- Generating 3+ substantial code artifacts (apps, components, modules) in one workflow\n- Each artifact needs 500+ lines\n- Single-call approach produces undersized results\n\n## Pattern\n\n```\nCall 1 (lightweight, 2min timeout):\n  \"Generate N ideas with names, descriptions, feature",
  "cost": {
    "context_tokens": 328
  }
}

Fetch it by URL: GET /api/v1/registry/kjuhwa-skills-hub-split-llm-codegen-into-idea-plus-per-item/manifest?version=1.0.0

Reviews

Star ratings from people who tried it. One review per account; edit yours any time.

No reviews yet. Install it, try it, and be the first to rate it.