Project Starter is a workflow plugin that equips coding agents with specialized subagents, reusable skills, commands, hooks, and output styles for software projects. It is for developers who want structured assistance with tasks such as planning, debugging, code review, security, and automation across several coding-agent platforms. Catalogue add-ons are components of its own workflow.
Getting it into your agent
One page per mod, every tool's command on it. A separate URL per tool would split the same page into five that compete with each other.
git clone --depth 1 https://github.com/CloudAI-X/claude-workflow-v2Wrote this? Show the measurements
A badge with what this costs and how it scanned, read live from this page, so it follows the numbers instead of freezing them. Markdown for a README, HTML for a documentation site or a project page.
[](https://agentmods.dev/commands/cloudai-x/claude-workflow-v2/refactor-guided)<a href="https://agentmods.dev/commands/cloudai-x/claude-workflow-v2/refactor-guided"><img src="https://agentmods.dev/badge/commands/cloudai-x/claude-workflow-v2/refactor-guided/github.svg" alt="Measured on agentmods" height="20"></a>Or the 80×15 button, for a site that already has a row of RSS and ATOM ones. Only the verdict fits; the numbers stay here.
<a href="https://agentmods.dev/commands/cloudai-x/claude-workflow-v2/refactor-guided"><img src="https://agentmods.dev/badge/commands/cloudai-x/claude-workflow-v2/refactor-guided.svg" alt="Reviewed on agentmods" width="80" height="20"></a>What it costs to keep this loaded
Counted locally with the o200k_base tokenizer, which is exact for GPT models; Claude uses its own tokenizer and its counts differ. Treat this as one consistent yardstick across the catalogue rather than a bill. Prices are per million input tokens.
| Model | Per session | Once invoked |
|---|---|---|
| Fable 5.1 | $0.00021 | $0.00989 |
| Opus 5.5 | $0.00008 | $0.00396 |
| Sonnet 5.5 | $0.00004 | $0.00198 |
| Haiku 4.5 | $0.00002 | $0.00099 |
Grade A, and why
refactor-guided scanned grade A with 0 findings against 26 rules in 11 categories — prompt injection, anti-refusal, data exfiltration, privilege escalation, supply chain, agent snooping, system-prompt leakage, SSRF and excessive agency — measured 18d ago.
A static scan of the body, not an audit. Every finding is printed with the line that produced it so you can judge whether it matters here. A mod is markdown that instructs an agent; that is exactly why what it instructs is worth reading.
Nothing flagged
None of the 26 patterns this scan looks for appear in this file: no shell pipes, no recursive deletes, no credential paths, no hidden text, no instruction-override or anti-refusal phrasing, no agent-config snooping. That is not a guarantee, it is the absence of the things that are checkable.
How it starts
The opening of the file, as written. The whole thing — 118 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Guided Refactoring
Perform safe, incremental refactoring with test verification at every step.
Context
- Recent changes: !
git diff --name-only HEAD~3 2>/dev/null | head -10 - Current branch: !
git branch --show-current - Working tree status: !
git status --short - Project root files (infer the test framework from these): !
ls -a - Existing tests: !
git ls-files 2>/dev/null | grep -i -e test -e spec | head -10
Safety Rules (NON-NEGOTIABLE)
- NEVER refactor AND add features simultaneously. Each change is purely structural.
- NEVER proceed without passing tests. If tests fail after a refactoring, revert immediately.
- ONE refactoring at a time. Each commit is a single, atomic refactoring step.
- Preserve external behavior. Inputs and outputs must remain identical.
- Ensure clean working tree before starting. If there are uncommitted changes, stop and ask the user — do not stash or commit their work yourself.
Workflow
Phase 1: Analyze Scope
- Identify the target file(s) or directory from
$ARGUMENTS - Map all dependents — files that import/use the target
- Map all dependencies — files that the target imports/uses
- Assess blast radius: how many files could be affected?
- Report scope summary before proceeding:
## Refactoring Scope - Target: [file/directory] - Files affected: [count] - Dependencies: [list] - Dependents: [list] - Risk level: [low/medium/high]
Phase 2: Ensure Test Coverage
- Check if tests exist for the current behavior of the target
- Run existing tests to confirm they pass (this is your safety net)
- If test coverage is insufficient:
- Write tests that capture the current behavior BEFORE refactoring
- Run them to confirm they pass
- Commit these tests separately:
test: add coverage for [target] before refactoring
- Record baseline test results (pass count, time)
Phase 3: Refactor Incrementally
For each refactoring step:
- Describe the specific refactoring about to be applied (e.g., "Extract method X from class Y", "Rename variable A to B", "Move function to separate module")
- Apply the single refactoring change
- Run tests immediately after applying
- If tests PASS:
- Commit with message:
refactor: [specific change description] - Move to next refactoring step
- Commit with message:
- If tests FAIL:
- Revert the change:
git checkout -- ., and delete any new files this step created (they are untracked, so checkout leaves them behind) - Analyze why it failed
- Try a different approach or break it into smaller steps
- Document what went wrong
- Revert the change:
What this file has done since we first saw it
Hashed on every crawl. A supply-chain change to an agent config is a question of when, not whether, so the history is kept rather than the latest state alone.
- 18d ago Changed d4bde545c1a0
- 1mo ago First seen · 118 lines · 21 tokens per session scan A 4afd3b792eaf
refactor-guided is a command published in the GitHub repository CloudAI-X/claude-workflow-v2 (1,420 stars, last pushed 15d ago), licensed MIT. It adds 21 tokens to every session and 989 once invoked, about $0.0001 per session on Opus 5.5. A static security scan graded it A with 0 findings. No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-30.
Other commands, from other repositories
full-review
Orchestrate comprehensive multi-dimensional code review using specialized review agents across architecture, security, performance, testing, and best practices.
kiro-validate-impl
Command "kiro-validate-impl" from gotalab/cc-sdd, covering implementation validation, core task, execution steps, 1. detect validation target and 2. load context.
final-review
Command "final-review" from adriannoes/awesome-agentic-ai, covering final review - comprehensive pr review & testing, step 0: determine review pass, step 1: create or update the pr, step 2: launch three review agents in parallel and agent 1: codebase consistency reviewer.
arch-boundaries
Install mechanically-enforced architecture boundaries that fail CI on spaghetti imports.
validate-parallel
Run assert and review-code commands in parallel to reduce validation time by 50%.
ijfw-audit
Run the IJFW audit gate for the current workflow phase. Usage: /ijfw-audit [phase name].