spec-shaping_skill

This skill helps shape product ideas into actionable specs and sprint plans, clarifying scope and enabling focused delivery.
  • TypeScript

0

GitHub Stars

2

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 nweii/agent-stuff --skill spec-shaping

  • changelog.md1.6 KB
  • SKILL.md7.4 KB

Overview

This skill shapes product ideas into clear specs and actionable sprint plans. It supports in-depth interviewing to narrow vague problems and a sprint-breakdown mode that produces atomic tasks, demoable sprints, and verifiable validation criteria. Use it to turn a fuzzy plan into a repeatable engineering roadmap.

How this skill works

In Interview mode, the skill asks targeted, non-obvious questions to narrow the problem, surface hidden dependencies, and confirm the true user need before writing or updating a spec file. In Sprint Breakdown mode, it reads an existing spec, produces a task registry with stable IDs and dependencies, groups tasks into demoable sprints, and annotates each task with concrete validation criteria and traceability back to spec sections.

When to use it

  • You have a rough product idea that needs precise scoping
  • You want to validate that a problem is well-framed before building
  • You need to decompose a ready spec into atomic tasks and sprints
  • You want to surface hidden technical or UX risks early
  • You are preparing a spec file and want consistent document hygiene

Best practices

  • Search existing specs before creating new files; update when the core problem identity remains the same
  • Name files with a semantic title and date, include frontmatter status and last-modified date
  • Keep a Context header and a Changelog section; log only significant scope changes
  • Make tasks atomic and include explicit validation criteria for each task
  • Order tasks so dependencies are done first and reserve 10–20% of sprint capacity for research/spikes

Example use cases

  • Interview a PM who proposes “build a calendar” and narrow it to a single, testable feature spec
  • Take a completed spec and produce T1..Tn task registry with dependencies and sprint groupings
  • Refine a pivoted product direction by updating the existing spec file and adding a changelog entry
  • Run a subagent review on a complex sprint plan to catch circular dependencies or oversized tasks
  • Convert UX mockups and acceptance criteria into demoable first-sprint deliverables

FAQ

If the core problem identity is the same, update the existing spec and add an update date; create a new file only when the problem definition fundamentally changes.

How detailed should validation criteria be?

Validation must be concrete and verifiable — API status and payload checks, Storybook renders, or user-visible demo steps. If you can’t state a clear pass/fail, refine the task further.

Do you provide time estimates or resource assignments?

No. This skill avoids time estimates and resource allocation unless you explicitly request them with sufficient context.

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