agentsclimarketplace

Batch orchestration

Skill subhansh-dev/agent-maxxing/workflows/batch-orchestration

95+ agent skills, 19 UI components, 7 system prompts from Claude Fable 5, GPT-5.5, Gemini CLI & more. Self-fine-tune your agent: paste one prompt and it reads every skill, internalizes patterns, and becomes elite. Works with Claude Code, Codex, Cursor, OpenCode + 60 more agents.

Install
npx -y skills add subhansh-dev/agent-maxxing --skill batch-orchestration

Assembled from the repository path, not quoted from the project. Check it against their README if it does not work.

2 things to look at

  • 29 days oldThe repository was created 29 days ago. New is not bad, but a brand new repository carrying a familiar-sounding name is the shape a typosquat arrives in, and there has been no time for anyone else to find a problem with it.
  • 2 stars2 stars. Stars are a popularity signal and not a quality one, but at this level it is likely that nobody has read this closely except its author, and you would be relying on your own review.

What its author says it does

Copied from the file, not written here

Orchestrate large parallelizable changes by decomposing into independent units and spawning parallel worker agents in isolated worktrees.

SKILL.md

2.1 KB, 462 tokens by cl100k_base, as published. Nobody here has run it

Batch Orchestration — Parallel Work

Extracted from Claude Code's bundled batch skill.

Phase 1: Research and Plan

  1. Understand the scope. Launch subagents to deeply research what the task touches. Find all files, patterns, and call sites that need to change.

  2. Decompose into independent units. Break the work into 5-30 self-contained units. Each unit must:

    • Be independently implementable in an isolated git worktree
    • Be mergeable on its own without depending on another unit's PR
    • Be roughly uniform in size

    Scale the count to the actual work: few files → closer to 5; hundreds → closer to 30. Prefer per-directory or per-module slicing.

  3. Determine the e2e test recipe. Figure out how a worker can verify its change actually works end-to-end:

    • Browser automation for UI changes
    • CLI verifier for terminal changes
    • Dev-server + curl for API changes
    • Existing e2e/integration test suite

  4. Write the plan with:

    • Summary of research findings
    • Numbered work units (title, files, description)
    • E2e test recipe
    • Worker instructions template

Phase 2: Spawn Workers

Spawn one background agent per work unit. All agents use isolation: "worktree" and run_in_background: true. Launch them all in a single message block.

Each agent's prompt must be self-contained:

  • Overall goal
  • Specific work unit description
  • Files/directories to modify
  • E2e test recipe
  • The exact changes to make

Phase 3: Collect and Merge

After all workers complete:

  1. Review each worktree's changes
  2. Verify e2e tests pass for each
  3. Merge worktrees back into the main branch
  4. Resolve any conflicts
  5. Run a final integration test

Rules

  • Workers must not modify files outside their assigned scope
  • If a worker fails, diagnose and retry with more context
  • Never skip e2e verification — unit tests aren't enough
  • Keep worker prompts concise but complete

Keep looking

Skills are one crate of 328,083. Ordering is by how many stacks a row turns up in, so the top of any crate is what has actually been picked rather than what has the most stars.