dev-workflow
End-to-end development workflow for repositories that use phase planning, red-team review, implementation, pull requests, independent review, and deployment…
Skill by itsAR-VR
Install
Works with Claude Code, Claude apps, Codex, Cursor, Gemini CLI, GitHub Copilot, VS Code, OpenCode, Goose, Windsurf. Read the repository before you give any tool access to your data.
npx agentmarket install skill/itsar-vr-dev-workflow
Asks which of your tools to add it to (Claude Code, Codex, Cursor, VS Code, Windsurf, Gemini CLI, Claude Desktop) and shows the change first. See https://agentmart.dev/cli.
Get the weekly diff
The weekly diff: one email a week with new agent tools, notable updates and grade changes, plus security alerts for tools you follow. Double opt-in, one-click unsubscribe.
About
Use cases: productivity
Reviews
Write a reviewNo reviews yet. Installed it? Reviews come from verified installers only. How reviews work
Permissions
What this skill can do and what it asks for, from its SKILL.md. Checked Oct 25, 2026.
- Tools
- No tools found
- Secrets
- None asked for
- Install scripts
- None found
Network hosts it talks to 1
codeload.github.com· found in install command
Files it touches
~/.claude/skills/dev-workflow
What's inside
- Skill name
dev-workflow- Description
- End-to-end development workflow for repositories that use phase planning, red-team review, implementation, pull requests, independent review, and deployment proof. Use before nontrivial coding, debugging, refactoring, or release work.
- License
- MIT
- Instructions
- 47 lines in SKILL.md · view on GitHub
Files (1)
- SKILL.md
SKILL.md (first lines)
# Development Workflow Use this workflow when a repository does not provide stricter local instructions. Repository-level `AGENTS.md` or equivalent instructions always win. ## Rules - Treat the default branch as read-only. Ship through a task branch and pull request. - Read the files, nearby patterns, tests, and current git state before editing. - Preserve dirty work. Fetch for freshness and use a clean worktree when the active checkout is unsafe. - Start nontrivial work with a versioned plan that names the goal, scope, risks, verification, and stop conditions. - Red-team the plan with `deep-sweep` before implementation. - Run impact analysis before changing shared symbols. If no code-graph tool exists, search every usage manually. - Build the smallest complete change with `deep-build`; do not mix unrelated cleanup into the diff. - Test the failing behavior first for bug fixes, then run focused and regression checks. - Use an independent reviewer for anything user-facing or release-bound. The producer does not grade its own work. - Do not merge until required checks, review findings, and deployment or runtime proof are complete. - Never print, commit, or publish credentials, private customer data, raw sessions, or local authentication state. ## Pipeline 1. Refresh read-only context: branch, status, remotes, default branch, open pull requests, and issue state. 2. Write the plan and success criteria. 3. Run `deep-sweep`; amend the plan for valid findings. 4. Create a task branch from the current default branch. 5. Run impact analysis for each symbol before editing it. 6. Implement with `deep-build` and keep the diff surgical. 7. Run focused tests, type checks, builds, and end-to-end checks proportional to risk. 8. Inspect the final diff and run a secret/private-data scan. 9. Open a pull request with source context, validation, risks, and proof. 10. Resolve independent review findings and re-run affected checks. 11. Merge only after gates pass, then verify the real deployed or runtime surface. ## Done Means Proven A task is complete only when the requested outcome is visible on the surface that matters. A local build does not prove a deployment, a generated file does not prove it renders, and an open pull request does not prove the change shipped. Record: - branch and commit - pull request and check state - commands run and their results - runtime, browser, or artifact proof - remaining human or external gates If a repository has no formal planning or deployment system, keep the same sequence and use the lightest local equivalent.
From itsAR-VR/public-skills, MIT.
More from itsAR-VR/public-skills
- 0.474score
browser-harness
contributor
Default browser automation in the user's real, already-logged-in Chrome.
SkillCFreshness: CClaude Code +9 - 0.474score
gitnexus
contributor
Self-bootstrapping codebase knowledge graph.
SkillCFreshness: CClaude Code +9 - 0.474score
verify
contributor
Post-execution quality gate combining goal-backward verification and 8-track cleanup into one coherent skill.
SkillCFreshness: CClaude Code +9 - 0.474score
adcrate-cold-audience
itsAR-VR
Use when writing, rewriting, reviewing, or generating any copy or content for a cold or low-context audience that does not already know, trust, or actively…
SkillCFreshness: CClaude Code +9 - 0.467score
goal-post
itsAR-VR
Turn an outcome into a machine-checkable goal contract with actor, verifier, stop conditions, evidence, and human gates.
SkillCFreshness: CClaude Code +9 - 0.467score
deep-clean
itsAR-VR
Remove accidental complexity after a working implementation without changing behavior.
SkillCFreshness: CClaude Code +9 - 0.467score
brave-search
itsAR-VR
Web search and content extraction via Brave Search.
SkillCFreshness: CClaude Code +9 - 0.467score
deep-build
itsAR-VR
Execute an approved implementation plan in small verified slices.
SkillCFreshness: CClaude Code +9 - 0.467score
deep-sweep
itsAR-VR
Red-team a plan before implementation by checking assumptions, scope, failure modes, security, dependencies, and proof.
SkillCFreshness: CClaude Code +9
Signals
- Stars
- 0
- Forks
- 1
- Open issues
- 1
- Last commit (this folder)
- Jul 14, 2026 (3 months ago)
From GitHub (fetched Oct 24, 2026).
Details
- License
- MIT
- Repository
- github.com/itsAR-VR/public-skills
- First seen
- Oct 24, 2026
- Last checked
- Oct 24, 2026
- Source
- GitHubgithub-graphql
More from itsAR-VR
- 0.474score
adcrate-cold-audience
itsAR-VR
Use when writing, rewriting, reviewing, or generating any copy or content for a cold or low-context audience that does not already know, trust, or actively…
SkillCFreshness: CClaude Code +9 - 0.467score
deep-build
itsAR-VR
Execute an approved implementation plan in small verified slices.
SkillCFreshness: CClaude Code +9 - 0.467score
brave-search
itsAR-VR
Web search and content extraction via Brave Search.
SkillCFreshness: CClaude Code +9 - 0.467score
deep-clean
itsAR-VR
Remove accidental complexity after a working implementation without changing behavior.
SkillCFreshness: CClaude Code +9 - 0.467score
deep-sweep
itsAR-VR
Red-team a plan before implementation by checking assumptions, scope, failure modes, security, dependencies, and proof.
SkillCFreshness: CClaude Code +9 - 0.467score
goal-post
itsAR-VR
Turn an outcome into a machine-checkable goal contract with actor, verifier, stop conditions, evidence, and human gates.
SkillCFreshness: CClaude Code +9
Cite this page
dev-workflow skill. Agent Mart. https://agentmart.dev/skills/itsar-vr-dev-workflow (last updated 2026-10-24)