autospec-release

Compute deterministic PASS/PARTIAL/FAIL release verdicts from QA evidence.

3|1|Updated Apr 29, 2026
One-click install
npx skills add https://github.com/berlinguyinca/autospec --skill autospec-release
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: autospec-release
Source: https://github.com/berlinguyinca/autospec/tree/main/skills/autospec-release
Command: npx skills add https://github.com/berlinguyinca/autospec --skill autospec-release

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

autospec-release determines whether a repository is genuinely ready to ship by coordinating specs, docs, implementation state, automated tests, live QA proof, and legacy cleanup into a single PASS/PARTIAL/FAIL verdict.

Core Features & Use Cases

  • Release readiness pipeline: Executes a full gate that starts with preflight context, then runs sweep, review, classification/running of remaining work, tests, QA proof, and legacy cleanup.
  • Live-app QA proof enforcement: Requires interaction-level evidence via /autospec-qa and uses .autospec/qa-verdict.json to compute the final verdict for the current HEAD.
  • Deterministic verdict computation: Uses scripts/compute-release-verdict.sh as the canonical source of truth, producing a consistent outcome and separating proven evidence from NOT TESTED gaps.
  • Release loop with remediation: Iteratively re-runs the smallest needed stage(s) until the release gate is satisfied or iteration limits/operator stop conditions are reached.

Quick Start

Run autospec-release to produce a single release verdict for your current repo state.

Frequently Asked Questions about autospec-release

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

FAQPage Schema
How do I determine if my repository is ready to ship with a release readiness gate?

A release readiness gate verifies repository readiness to ship by coordinating specs, docs, implementation state, automated tests, and live QA proof into a single PASS, PARTIAL, or FAIL verdict.

How does live QA proof enforcement work during a release pipeline?

Live QA proof enforcement requires interaction-level evidence via the /autospec-qa stage, writing results to .autospec/qa-verdict.json to compute the final deterministic release verdict for the current HEAD.

What is documentation drift and how do I check for it before a release?

Documentation drift occurs when project docs fall out of sync with code. A release readiness pipeline runs a sweep stage to detect drift and ensure documentation matches the current implementation state before shipping.

Can I use a deterministic script to compute a release verdict from QA test gates?

Yes, the scripts/compute-release-verdict.sh script acts as the canonical source of truth, deterministically computing the release verdict from .autospec/qa-verdict.json to separate proven evidence from NOT TESTED gaps.

Does autospec-release work with any Git repository or are there specific prerequisites?

It applies exclusively to existing autospec-managed repositories, requiring the full release pipeline stages including mandatory /autospec-qa to run before computing the release verdict.

What happens if a release readiness gate fails or returns a PARTIAL verdict?

The release loop iteratively re-runs the smallest needed pipeline stages for remediation until the release gate is satisfied, or until iteration limits and operator stop conditions are reached.