Generate a tasks.md file with T-XXX IDs, AC coverage, and parallelization markers.
npx skills add https://github.com/gaabscps/ai-squad --skill task-builder-gaabscps
Please help me install this Agent Skill. Skill: task-builder Source: https://github.com/gaabscps/ai-squad/tree/main/squads/sdd/skills/task-builder Command: npx skills add https://github.com/gaabscps/ai-squad --skill task-builder-gaabscps
Task Builder solves the manual, error-prone process of turning an approved Spec and Plan into a precise, task-level plan (tasks.md) that can be executed in phases with AC coverage and [P] parallelization markers.
Invoke /task-builder FEAT-NNN after Spec and Plan are approved to generate the initial tasks draft.
High-intent search queries and answers about installing and using this skill.
To decompose an approved spec into a task graph, you can automate the creation of a granular tasks.md file that maps T-XXX tasks with acceptance criteria coverage and parallelization markers. This enforces preconditions and translates plans into verifiable execution.
Generating a tasks.md file with AC coverage is best handled by an automated workflow that decomposes approved specs into atomic tasks with US references and optional [P] parallelization markers. This ensures verifiable phased execution.
Vertical-slice decomposition with Spec Kit patterns works by interactively translating approved specs and plans into a detailed tasks.md file. It assigns T-XXX IDs and maps acceptance criteria coverage for each slice.
Yes, you need an approved plan and spec before generating task-level execution plans. The workflow validates preconditions and enforces approval states before producing any task drafts, though a PM-mode bypass logic is supported.
Yes, you can use task decomposition for phased project management by generating a tasks.md file with optional [P] parallelization markers. This allows verifiable, phased execution of the approved spec and plan.
Automated task scoping enforces preconditions to ensure specs and plans are fully approved before generating drafts. This prevents errors from manual translation and maintains valid approval states through atomic session state writes.
yi-john-huangGenerate EARS-formatted requirements documents with FRs, NFRs, and acceptance criteria.
wtahValidate a single FR-, AI-, or UJ-item against a running application.
Covcloud-LLCCoordinate multi-session AI development workflows across git repositories.
DoubleslashSEStructure tasks with IDs, dependencies, verification steps, and conventional commits.
