tech-spec

Transform approved PRDs into sprint-based technical specifications and runtime bundles.

2|Updated Apr 25, 2026
One-click install
npx skills add https://github.com/iurykrieger/claude-yoke --skill tech-spec-iurykrieger
Or copy as Structured Prompt for Agent▼
Please help me install this Agent Skill.
Skill: tech-spec
Source: https://github.com/iurykrieger/claude-yoke/tree/main/skills/tech-spec
Command: npx skills add https://github.com/iurykrieger/claude-yoke --skill tech-spec-iurykrieger

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

Phase 2 — Technical specification planning becomes automated: turning approved PRDs into structured sprint specs, DoDs, and per-sprint runtime bundles. The skill orchestrates a three-stage blueprint (LLM drafts the sprint index, bash scaffolds empty per-sprint files, LLM fills each sprint) and stores artifacts under .yoke/specs/ and .yoke/sprints/. It pauses at Trigger 2 until approval, ensuring a controlled, auditable transition to Phase 3.

Core Features & Use Cases

  • Drafts a sprint index from PRD to guide sprint planning and task allocation.
  • Scaffolds empty per-sprint files deterministically via Bash, seeded with frontmatter skeletons.
  • Fills sprint files with objective, DoD, tasks, functional criteria, and sensors using LLM prompts, one sprint at a time.
  • Enforces Trigger 2 approval before transitioning to Phase 3 and marks artifacts as approved.

Quick Start

Initiate the workflow by providing an approved PRD to /yoke:tech-spec to generate the sprint index, scaffold per-sprint files, and iteratively fill them.

Frequently Asked Questions about tech-spec

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

FAQPage Schema
How do I convert an approved PRD into sprint-ready technical specifications?▼

To convert an approved PRD into sprint-ready technical specifications, the skill drafts a sprint index using an LLM, scaffolds empty sprint files via Bash, and fills each file with objectives, tasks, and DoDs before requiring approval.

How does generating per-sprint runtime bundles from PRDs work?▼

Generating per-sprint runtime bundles works through a three-stage workflow: LLM drafts the sprint index, Bash deterministically scaffolds empty per-sprint files, and the LLM fills each file with strict binary acceptance criteria.

Can I use this skill to enforce an approval workflow before starting development sprints?▼

Yes, you can enforce an approval workflow because the skill pauses at Trigger 2, requiring explicit approval before all sprints transition to Phase 3, ensuring a controlled and auditable development handoff.

What is the best way to structure sprint tasks with strict binary acceptance criteria?▼

The best way to structure sprint tasks with strict binary acceptance criteria is by having the LLM populate each sprint file with functional criteria, sensors, and tasks gated by a Trigger 2 approval.

Where are technical specification and sprint artifacts stored during this process?▼

Technical specification and sprint artifacts are stored as markdown files under .yoke/specs/<slug>.md and .yoke/sprints/<slug>-sNN>.md, keeping all sprint planning outputs organized in your workspace.