Agent skill · AI & Agents

openspec-plus-apply

MANDATORY skill that activates whenever the OpenSpec apply phase begins. Triggers: /opsx-apply runs, the openspec-apply-change vanilla skill is referenced or active, `openspec instructions apply` is invoked, or the user asks to implement, apply, execute, or build out an OpenSpec change ('implement the change', 'apply tasks', 'execute change', 'build out the change'). Takes over only vanilla step 6 (implementation loop) and emulates step 7 output (final status).

sudokar135★ · +60/wk · 1 repos on radarProfile →
claude-codecopilotwindsurfMIT
Install
npx skills add sudokar/openspec-plus --skill openspec-plus-apply --agent claude-code

Same command for any agent — swap --agent for codex, cursor, copilot.

Facts
Files in the skill folder: 5
SKILL.md size: 28 KB
Bundled scripts: none
Version: 1.2.1
Path: skills/openspec-plus-apply/SKILL.md
Open the folder on GitHub →
Where it comes from
Stars: 142 · +7 this week
Read our review of the source →

Weekly change comes from our own snapshots, not the repository page — it measures attention, not adoption.

Review
written from the skill's own SKILL.md · Aug 5, 2026

What it does

The skill activates during the OpenSpec apply phase to take over only vanilla step 6 (implementation loop) and emulate step 7 output (final status). It wraps the task implementation process with subagent orchestration, two-stage reviews (spec-compliance then code-quality), and strict test-driven development discipline. It supports conditional parallelism based on a dependency-validated slice graph, per-slice gates including lint/format/test checks, and a final gate that emulates vanilla step 7. It never modifies vanilla steps 1-5 or 7, never commits code, and never reads affected source files into the main agent context; subagents read files via provided paths.

How it works

  • Phase 0 (Pre-Flight) verifies inputs (proposal.md, spec.md, design.md, tasks.md) and artifacts produced by vanilla steps 2-4. It identifies required pre-check commands (lint/format/test, etc.) from project standards and sets a mandatory Mode Question to choose subagent mode.
  • Phase 1 builds a dependency graph of slices from tasks.md and design/spec cues, determines parallelizable groups, and asks about parallelism if independent slices exist.
  • Phase 2 per-slice execution loops dispatch implementer and reviewer subagents for each slice (or parallel group) in a mandated order. Before each dispatch, it requires reading the corresponding subagent prompts verbatim from dedicated prompt files, filling in placeholders such as {SLICE_NUMBER}, {SLICE_NAME}, {TASKS_TEXT}, {SPECS_TEXT}, {DESIGN_PATH}, {AFFECTED_FILES}, {PROJECT_STANDARDS_PATHS}, {LINT_CMD}, {FORMAT_CMD}, {TEST_CMD}, {OTHER_CHECKS}, {WORKING_DIR}.
  • Implements a 2-stage review: spec-compliance reviewer first, then code-quality reviewer, with per-slice gates and a maximum of three fix cycles. It maintains phase-based gating and never skips or reorders phases.
  • Phase 3 emulates the final gate/output (vanilla Step 7) after all slices succeed, without performing the actual vanilla Step 7.

When to use it

Use when initiating an OpenSpec change that requires overridden Step 6 execution with subagent orchestration, two-stage reviews, and strict adherence to the OpenSpec-plus TDD process. It activates automatically on triggering OpenSpec apply phase or when specific prompts such as openspec instructions apply are invoked.

What it can touch

  • The workflow touches the following via subagents: affected file paths (AFFECTED_FILES), design path (DESIGN_PATH), project standards paths (PROJECT_STANDARDS_PATHS), and shell commands for lint/format/test as discovered in Phase 0 (LINT_CMD, FORMAT_CMD, TEST_CMD, OTHER_CHECKS).

Caveats

  • This skill emulates Step 7 output and does not invoke vanilla Step 7 directly. It never commits code. It strictly follows the Phase 0 mode question and subagent dispatch protocol. It prohibits reading source code files into the main agent context; subagents read affected files from provided paths. It enforces a rigid sequence of phases and review cycles with a 3-fix-cycle limit before escalation.
From the SKILL.md

# OpenSpec Plus Apply ## Mission Wrap vanilla `/opsx-apply` step 6 (task implementation loop) with subagent orchestration, two-stage review, strict TDD via `openspec-plus-tdd`, four implementation principles, conditional parallelism over a dependency-validated graph, per-slice + final lint/format/test/other gates, never-ignore-failures rule. Vanilla owns: * Step 1 — Select change * Step 2 — `opens

More from openspec-plus
All skills →
About this skill
What does the openspec-plus-apply skill do?

MANDATORY skill that activates whenever the OpenSpec apply phase begins. Triggers: /opsx-apply runs, the openspec-apply-change vanilla skill is referenced or active, `openspec instructions apply` is invoked, or the user asks to implement, apply, execute, or build out an OpenSpec change ('implement the change', 'apply tasks', 'execute change', 'build out the change'). Takes over only vanilla step 6 (implementation loop) and emulates step 7 output (final status).

How do I install it?

Run `npx skills add sudokar/openspec-plus --skill openspec-plus-apply --agent claude-code` — it drops the skill into your project so the agent can pick it up. Swap the --agent value for codex, cursor or copilot if you use one of those.

Where does this skill come from?

From sudokar/openspec-plus, a repository with 142 stars. We read it straight from the repository tree rather than a submitted listing, so what you see here is what is actually published.

Is a popular skill a good skill?

Not necessarily. Stars measure attention, not adoption — a repository can trend for a week and be abandoned. That is why we show the weekly change from our own snapshots next to the total, instead of a single flattering number.

Keep going