All skills
simota avatar

/sherpa

@35ffd55
by shingo imotasimota/agent-skills85 stars
15

Guiding workflows by decomposing complex tasks (Epics) into Atomic Steps under 15 minutes each, with progress tracking and drift prevention. Use when complex decomposition is needed.

Use this Skill: https://skilld.dev/gh/simota/agent-skills/sherpa

This session only. Nothing lands on disk.

referencetask-decomposition-anti-patterns.md

≈448 tokens on demand. Your agent reads this file only when SKILL.md points to it.

Task Decomposition Anti-Patterns

Purpose: Use this file to validate decomposition quality during MAP, especially when steps are too large, too early, or structurally misleading.

Contents

  • TD-01 to TD-07
  • Decomposition signals
  • Vertical vs horizontal slicing
  • WBS checklist

Anti-Patterns

ID Anti-pattern Symptom Fix
TD-01 God Task many steps still exceed 15 min; no clear test or commit point keep decomposing to atomic steps
TD-02 Premature Decomposition detailed plan before the requirements are clear investigate first, then decompose gradually
TD-03 Skill-Based Splitting “coding / testing / docs” slices create silos prefer functional or component slices
TD-04 Template Addiction every story gets the same boilerplate breakdown let tasks emerge from the actual work
TD-05 Missing 100% Rule hidden work appears later explicitly check setup, tests, docs, and config
TD-06 Dependency Blindness parallel work later proves serial create a dependency graph during decomposition
TD-07 Over-Decomposition 20+ tasks or status work dominates real work regroup and reduce management overhead

Decomposition Signals

Split Further When

  • a step exceeds 15 min
  • the test method is unclear
  • no clean commit point exists
  • multiple files or modules are touched
  • the step contains implicit sequencing such as “do X, then Y”

Stop Splitting When

  • work is under 5 min
  • the sub-step cannot be verified on its own
  • tracking cost exceeds execution cost
  • the current list already exceeds about 20 items

Sweet Spot

  • Epic -> 3-5 Stories
  • Story -> 3-8 Tasks
  • Task -> 2-4 Atomic Steps
  • Atomic Step -> do not decompose further

Source: SKILL.md on GitHub

No alerts13d5 checks · Risk SAFE
  • Gen Agent Trust Hub13d

    The skill provides a robust framework for decomposing complex tasks into manageable atomic steps, featuring extensive protocols for risk management and focus protection. No security issues were identified.

  • Socket13d

    No alerts

  • Snyk13d

    Risk: LOW · No issues

  • Runlayer6mo

    1/11 files flagged

  • ZeroLeaks5mo

    Score: 93/100 · 2 sections analyzed

Signed by skilld at 35ffd55. This ties the file your Agent reads to that commit on GitHub. It does not review the instructions.

Last checked against GitHub 2 days ago.

Activeupdated 2 weeks ago

README badge

README badge for simota/agent-skills/sherpa