openspec-plus-apply

Orchestrates OpenSpec change implementation with subagent dispatch, two-stage review, and strict TDD.

1|Updated Oct 27, 2025
One-click install
npx skills add https://github.com/WndDnz/modelo-relatorio-unifei-ict --skill openspec-plus-apply-wnddnz
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: openspec-plus-apply
Source: https://github.com/WndDnz/modelo-relatorio-unifei-ict/tree/main/.opencode/skills/openspec-plus-apply
Command: npx skills add https://github.com/WndDnz/modelo-relatorio-unifei-ict --skill openspec-plus-apply-wnddnz

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve? Implementing an OpenSpec change by hand often means losing track of tasks, skipping tests, or letting unreviewed code slip through. This Skill takes over the implementation loop of the OpenSpec apply phase and enforces a disciplined, review-gated workflow so every task is implemented, tested, and verified before being marked complete. ## Core Features & Use Cases - Subagent Orchestration: Dispatches isolated implementer and reviewer subagents per slice, keeping the main agent's context clean by passing file paths instead of reading source code. - Two-Stage Review: Runs a spec-compliance review followed by a code-quality review for every slice, with a 3-cycle fix cap and escalation when artifacts are wrong. - Strict TDD Enforcement: Applies the openspec-plus-tdd per-test state machine (RED, VERIFY-RED, GREEN, VERIFY-GREEN, REFACTOR) one test at a time, with mandatory acceptance tests per Gherkin scenario. - Dependency-Aware Parallelism: Builds a slice dependency graph, validates file existence via git history, and optionally dispatches independent slices in parallel after user confirmation. - Use Case: After running /opsx-apply on a change, the Skill analyzes tasks.md, asks for execution mode, implements each slice under TDD with independent reviewers, runs lint/format/test gates, and emits the final implementation status. ## Quick Start Run /opsx-apply for your OpenSpec change and let this Skill take over the implementation loop, choosing subagent mode when prompted.

Frequently Asked Questions about openspec-plus-apply

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

FAQPage Schema
How do I implement an OpenSpec change with subagents?

Run /opsx-apply to load the change artifacts, then this Skill takes over the implementation loop. It asks for an execution mode, builds a slice dependency graph from tasks.md, and dispatches an implementer subagent plus two reviewer subagents per slice.

What is the difference between subagent mode and inline mode?

Subagent mode dispatches isolated implementer and reviewer subagents per slice, keeping the main agent's context free of source code. Inline mode runs the same TDD discipline and gates directly in the main agent as a fallback when subagent dispatch is unavailable.

Does this Skill enforce test-driven development?

Yes. It loads the openspec-plus-tdd skill and enforces one test at a time through the full RED, VERIFY-RED, GREEN, VERIFY-GREEN, REFACTOR cycle. Every Gherkin scenario must have a passing acceptance test, and skipping or disabling tests is forbidden.

Can slices of an OpenSpec change be implemented in parallel?

Yes, when slices have no dependency edges and disjoint affected-file sets. The Skill builds a dependency graph, validates file existence with git log, and asks for user confirmation before dispatching independent slices in parallel.

What happens when a reviewer keeps rejecting an implementation?

Each reviewer and gate has a 3-cycle fix cap. After three failed cycles the Skill stops, pauses the implementation, and suggests updating the spec, design, or tasks artifacts instead of attempting a fourth fix.

Does the Skill commit code or archive the change automatically?

No. It never commits code and never auto-triggers /opsx-archive or openspec-verify-change. After all gates pass it emulates the vanilla step 7 status output and only suggests archiving as the next manual step.