spae-build

Executes one atomic task from PLAN.md with test-driven implementation and state persistence.

1|Updated May 12, 2026
One-click install
npx skills add https://github.com/mystilleef/spae-framework --skill spae-build-mystilleef
Or copy as Structured Prompt for Agent▼
Please help me install this Agent Skill.
Skill: spae-build
Source: https://github.com/mystilleef/spae-framework/tree/main/skills/spae-build
Command: npx skills add https://github.com/mystilleef/spae-framework --skill spae-build-mystilleef

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill includes references (resource) components.

What problem does it solve? Multi-step AI coding workflows drift, hallucinate, and lose context between sessions. This Skill executes exactly one atomic task from a persisted plan, keeping implementation small, verifiable, and resumable by any agent harness. ## Core Features & Use Cases - Atomic task execution: Reads STATE.json and PLAN.md, implements exactly one task, and hands off to the check phase without claiming completion. - Test-driven workflow: Enforces failing-test-first development with exhaustive coverage of expected behavior, failure modes, and edge cases. - State persistence: Updates STATE.json phase and task status so any agent can resume the workstream from external artifacts. - Use Case: A team decomposes a feature into atomic tasks in PLAN.md; an agent runs the build phase to implement task T-002 with full tests, then hands off to the check phase for verification. ## Quick Start Run the build phase to execute the active atomic task from PLAN.md and update STATE.json for handoff to check.

Frequently Asked Questions about spae-build

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I execute an atomic task from a plan file with an AI agent?▼

Run the build phase after STATE.json shows phase build with an active task. The agent reads PLAN.md for the task's intent and acceptance criteria, writes failing tests first, implements minimal code, and sets the phase to check for verification.

What is test-driven development for AI coding agents?▼

Test-driven development for agents means writing a failing test for each acceptance criterion before any production code, then iterating until green. This Skill enforces exhaustive coverage of expected behavior, failure modes, and edge cases before handoff.

Can this workflow resume after an interrupted agent session?▼

Yes, state persists in STATE.json and PLAN.md rather than conversational memory. Any agent harness can re-read these artifacts and resume the workstream at the current phase and active task.

What happens when a task fails verification during build?▼

The agent returns to implementation and re-verifies until all criteria pass. It only fails the task for out-of-scope conditions like an infeasible criterion, plan defect, or broken external dependency, making no STATE.json write on failure.

When should I not use an atomic task execution workflow?▼

Avoid it for exploratory prototyping or one-off scripts where decomposition overhead exceeds the work itself. It fits structured, multi-task features where verifiability and resumability matter more than speed.