agentsclimarketplace

Task chunking

Skill nevitonsantana/adaptive-skills/skills/task-chunking

38 portable Agent Skills for disciplined, reviewable AI-assisted work across Codex, Claude Code, GitHub Copilot, and compatible agents.

Install
npx -y skills add nevitonsantana/adaptive-skills --skill task-chunking

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

  • no licenseNo license file was found in the repository. Code published without one is not open source by default, so using it at work is a question for whoever answers licensing questions where you are.
  • 1 stars1 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

Break oversized work into smaller, reviewable slices with explicit stop conditions and next-slice boundaries.

SKILL.md

2.6 KB, as published. Nobody here has run it

Overview

Use this skill when the task is too large, too fuzzy, or too entangled to execute safely as one block.

When to Use

  • the work enters as one oversized request
  • the first useful slice is still unclear
  • dependencies are making the task feel bigger than it should be
  • the team needs a clear stop condition before execution expands

When NOT to Use

  • the task is already a small, obvious slice
  • the real problem is feature definition rather than execution shape
  • the work needs macro gate or policy review instead of chunking

Core Moves

  1. Name the smallest useful slice that can stand on its own.
  2. Separate what belongs in this slice from what must wait.
  3. Identify the minimum dependencies needed for the current slice.
  4. Declare the stop condition for this round.
  5. Name the next slice only if it is already justified.

Optional Modules

  • Dependency thinning — reduce or defer cross-system coupling when the slice is too entangled.
  • Checkpoint insertion — add a short pause point when the work is likely to sprawl mid-round.
  • Handoff-ready slicing — make the current slice legible enough for another round or another operator to continue later.

Activation Triggers

  • Use dependency thinning when the task keeps expanding because of other systems or teams.
  • Use checkpoint insertion when the round is likely to grow past one clean execution pass.
  • Use handoff-ready slicing when another round, session, or agent may need to resume the work.

Expected Output

  • smallest useful slice
  • explicit in-slice versus later boundary
  • dependency notes for the current slice
  • stop condition for the round
  • optional next-slice marker

Verification

  • The current slice can be explained without listing the entire larger initiative.
  • The stop condition is explicit enough to prevent silent scope growth.
  • Dependencies kept for the current slice are truly necessary.
  • The next slice is not being smuggled into the current one.

Handoff Signals

  • The current slice is blocked by another ownership boundary.
  • The next slice would require a different skill or operating context.
  • Macro review or gate posture is needed before continuing.

Pairs Well With

  • workflow
  • feature-planning
  • checkpoint-review

Anti-patterns

  • Renaming a large task as a “slice” without reducing it.
  • Keeping hidden future work inside the current chunk.
  • Treating chunking as a substitute for product or architecture decisions.

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.