starwave-tasks_skill

This skill generates a precise, test-driven task plan for implementing a feature by converting design into incremental coding tasks with clear dependencies.
  • Python

16

GitHub Stars

1

Bundled Files

2 months ago

Catalog Refreshed

4 months ago

First Indexed

Readme & install

Copy the install command, review bundled files from the catalogue, and read any extended description pulled from the listing source.

Installation

Preview and clipboard use veilstrat where the catalogue uses aiagentskills.

npx veilstrat add skill arjenschwarz/agentic-coding --skill starwave-tasks

  • SKILL.md12.1 KB

Overview

This skill creates an actionable, test-driven implementation plan as specs/{feature_name}/tasks.md based on existing requirements.md and design.md. It validates feature discovery (folder or git branch), follows decision_log.md when present, and uses the rune CLI to write structured tasks ready for execution. After generating the plan, it asks: "Do the tasks look good?" to start the approval loop.

How this skill works

The skill inspects specs/{feature_name} for requirements.md and design.md and checks the current git branch to infer a feature name if needed. It generates a numbered checkbox task list that maps design elements to concrete coding tasks, assigns stream IDs, and encodes blocked_by dependencies. It uses TDD-first prompts for each implementation step and invokes rune to create specs/{feature_name}/tasks.md and any prerequisites.md when required.

When to use it

  • You have requirements.md and design.md for a feature and need a coding plan.
  • You want a test-driven, incremental checklist for converting design into code tasks.
  • You want tasks organized for parallel work streams with explicit dependencies.
  • You need a rune-created tasks.md file with structured metadata for agents.
  • You want to validate decision_log.md constraints are respected in tasks.

Best practices

  • Ensure specs/{feature_name}/requirements.md and design.md exist before running the skill.
  • Provide feature_name explicitly if the git-branch inference cannot find it.
  • Prefer small, test-first tasks that reference specific files and requirement IDs.
  • Keep tasks focused on code changes, tests, or file creation—no non-coding items.
  • Review and approve the generated tasks; the skill will iterate until you confirm.

Example use cases

  • Convert a new API design into a sequence of unit tests, implementations, and integration steps.
  • Break a frontend feature design into component tests, implementations, and wiring tasks across streams.
  • Add property-based tests where the design specifies such tests, with tests created before implementations.
  • Generate a prerequisites.md when manual setup (cloud keys, provisioning) is required before specific coding tasks.

FAQ

I will check the current git branch to infer the feature. If I still can’t find it I will ask: "I can't find this feature, can you provide it again? Based on the git branch, I think it might be {found_name}". If requirements.md or design.md are missing I will ask you to run the requirements and design skills first.

Will the skill write any production code?

No. The skill only creates the tasks.md (and prerequisites.md if needed). Actual implementation is handled by separate workflows or agents.

How are dependencies and parallel work organized?

Every task includes blocked_by arrays referencing task IDs for dependencies and a stream integer for parallel work. Streams group independent work by code locality or logical boundaries.

Built by
VeilStrat
AI signals for GTM teams
© 2026 VeilStrat. All rights reserved.All systems operational