develop · rohitg00 — Claude Code command: what it does, how to install
Runs structured feature development through research, planning, testing checkpoints, and strict verification before committing.
MIT ★ 2,966 updated 2026-09-29
/plugin marketplace add rohitg00/pro-workflow
/plugin install pro-workflow@pro-workflowWhat develop does
The command guides feature work through sequential gates. During the initial research phase, the agent inspects the codebase and grades readiness on a 100-point scale across five criteria: scope clarity, pattern familiarity, dependencies, edge cases, and test strategy. If the score stays below 70, the agent gathers more context instead of touching code.
Next comes an explicit plan covering modified files, new files, approach rationale, and verification steps. Implementation begins only when you respond with "proceed" or "approved". While coding, tests run after each file edit, with a forced pause for review every 5 changes.
The final review phase mandates a full re-read of every altered file. The agent greps for console.log, TODO tags, and hardcoded secrets, dropping any issue it cannot cite with an exact line number. Once verified, it formats a conventional commit and records project learnings in [LEARN] format.
What it helps with
- break large features into gated research and planning phases
- prevent untested code changes with per-file test runs
- catch forgotten debug logs and unverified warnings before git commit
- record implementation takeaways in a structured learning format
Example prompts for the agent
- /develop add GitHub OAuth login flow
- /develop migrate session store to Redis
- /develop implement paginated CSV export for invoices
How to install develop
1Plugin marketplace
/plugin marketplace add rohitg00/pro-workflow/plugin install pro-workflow@pro-workflow2Manually: put the files in place
.claude/commands/develop.mdCall it with /develop. For all projects use ~/.claude/commands/
Manually: put the files in place
~/.codex/prompts/develop.mdCall it with /prompts:develop. Codex marks this format deprecated in favour of skills
Manually: put the files in place
.github/prompts/develop.prompt.mdManually: put the files in place
.gemini/commands/develop.tomlGemini stores commands as TOML: the prompt text goes into the prompt field
Questions about develop
How to install develop in Claude Code?
Add the marketplace source with /plugin marketplace add rohitg00/pro-workflow, then run /plugin install pro-workflow@pro-workflow. You can also place the command file directly at .claude/commands/develop.md.
What does the develop command do?
It enforces a phased workflow (Research, Plan, Implement, Review & Commit) that requires passing confidence scores and explicit approval before any code changes occur.
When does the agent actually start writing code?
Implementation starts only after the research score reaches 70 or higher, the written plan is presented, and you explicitly type "proceed" or "approved".
Source text
develop.md · MIT
--- description: Build a feature using Research > Plan > Implement phases argument-hint: <feature description> --- # /develop - Multi-Phase Feature Development Build features through structured phases with validation gates. ## Feature: $ARGUMENTS ### Phase 1: Research Explore the codebase to understand the scope: 1. Find all relevant files and existing patterns 2. Check dependencies and constraints 3. Score confidence across 5 dimensions (0-100) **Scoring:** - Scope clarity (0-20): Know exactly what files change? - Pattern familiarity (0-20): Similar patterns exist? - Dependency awareness (0-20): Know what depends on changed code? - Edge cases (0-20): Can identify the edge cases? - Test strategy (0-20): Know how to verify? **Decision:** - Score >= 70 → Present research findings and move to Phase 2 - Score < 70 → Identify gaps, gather more context, re-score ### Phase 2: Plan Present a plan for approval: ```text PLAN: [Feature Name] Goal: [one sentence] Files to modify: 1. path/file.ts - [what changes] New files: 1. path/new.ts - [purpose] Approach: 1. [step with rationale] Risks: - [potential issue and mitigation] Test strategy: - [how to verify] ``` **Wait for "proceed" or "approved" before continuing.** ### Phase 3: Implement Execute the approved plan: 1. Make changes in plan order 2. Run tests after each file change 3. Pause for review every 5 edits 4. Run full quality gates at the end (lint, typecheck, test) ### Phase 4: Review & Commit Self-review with verification — every finding must be confirmed by reading the code. 1. **Read every changed file** — re-read each modified file in full 2. **Verify, don't assume** — for each potential issue, quote the exact line. If you can't quote it, drop the finding. 3. **Grep for problems** — run `grep` for console.log, TODO, hardcoded secrets, debug statements. Report only what grep finds. 4. **Never report unverified findings** — don't say "ensure X" or "consider Y". Either it's a confirmed problem with a file:line citation, or it's not worth reporting. 5. Present verified summary for final approval 6. Commit with conventional message ### Learning Capture After completing, ask: - What corrections were made during implementation? - Any patterns worth adding to LEARNED? - Format: `[LEARN] Category: Rule`
© Rohit Ghumare · MIT · rohitg00/pro-workflow
Similar slash commands
All slash commands for Agent setup: memory, context, orchestration →develop by other authors comes first.