sw-framework_skill

This skill helps you master SpecWeave framework conventions, increment lifecycle, and living docs synchronization for spec-driven development.
  • Python

1.1k

GitHub Stars

3

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 openclaw/skills --skill sw-framework

  • _meta.json279 B
  • MEMORY.md152 B
  • SKILL.md17.1 KB

Overview

This skill is an expert guide to the SpecWeave framework structure, rules, and spec-driven development conventions. It helps teams adopt increment-based workflows, enforce source-of-truth discipline, and configure hooks and plugins. Use it to learn best practices for spec.md/tasks.md formats, living docs sync, and file organization patterns.

How this skill works

The skill inspects increment folders under a single root-level .specweave/ directory and validates required files (spec.md, tasks.md, metadata.json). It enforces naming and lifecycle rules, suggests when to include plan.md, and verifies project/board frontmatter for proper living-docs placement. It also explains hook behavior and plugin detection phases to help configure automation and manual update responsibilities.

When to use it

  • When you need to create or review a new increment and ensure spec.md frontmatter is correct
  • When implementing the SpecWeave root-level repository layout and source-of-truth rules
  • When configuring or troubleshooting hooks (post-task-completion, pre-implementation, etc.)
  • When deciding whether a feature needs plan.md or can skip it (bugs, hotfixes)
  • When evaluating plugin suggestions and reducing context with a plugin set

Best practices

  • Keep a single .specweave/ at repository root to maintain one source of truth
  • Always edit source files in src/ and run install scripts to sync .claude/ artifacts
  • Name increments with descriptive kebab-case: ####-descriptive-kebab-case-name
  • Mark an increment DONE before starting the next; use /sw:status and /sw:close for enforcement
  • Place all AI-generated outputs inside the corresponding increment folder, never in repo root

Example use cases

  • Planning a small feature: create spec.md and tasks.md, skip plan.md for simplicity
  • Onboarding a new repo: enforce .specweave/ layout, populate CLAUDE.md and README.md
  • Hook tuning: enable post_task_completion sound and logs while leaving automatic task updates off
  • Plugin enablement: run plugin detection during init and first increment to reduce context size
  • Increment recovery: use /sw:close to move incomplete tasks to the next increment or shrink scope

FAQ

Every increment must include spec.md, tasks.md, and metadata.json; plan.md is optional for simple work.

Where must the .specweave/ folder live?

It must be at the repository root. Nested .specweave/ folders are not supported; for multi-repo projects use a parent folder with one .specweave/.

Can I edit installed files in .claude/?

No. Always edit the source in src/ and run install scripts. .claude/ is generated and will be overwritten.

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