How do I install the Flowforge skill?

Run npx skillmds@latest add dvcrn/flowforge in your terminal (requires Node.js), paste this page's agent-chat prompt into Claude, Cursor, or any MCP-connected agent, or download the SKILL.md file and copy it into your agent's skills directory.

Is Flowforge safe to use?

This skill has not completed SkillMD's automated safety review yet. Independent scanners report: SkillSpector: PASS, Skill Scanner: PASS. Capability flags: executes scripts. SkillMD never runs a skill's scripts for you; review the SKILL.md before installing.

Which AI agents work with Flowforge?

This skill is tagged as working with Claude Code, Claude.ai, OpenAI Codex. SKILL.md is an open format, so most agents that read a skills directory can load it too.

Is Flowforge free to use?

Yes. Installing skills from SkillMD is free, and the skill stays under its author's original license.

Who published Flowforge?

dvcrn (@dvcrn) published this skill. Their other Agent Skills are listed on their SkillMD profile.

Flowforge by dvcrn — Agent Skill

Name: Flowforge
Rating: 4.5 (4 reviews)
Author: dvcrn

Flowforge

Autonomous AI coding pipeline that breaks any task into a structured Spec → Plan → Code → QA loop and executes it via Claude Code. Use when: (1) starting any new feature, refactor, or bug fix, (2) given a GitHub issue to implement, (3) asked to "run FlowForge", "forge this", "plan and build", or "auto-implement". Routes ALL heavy work through Claude Code. Supports multi-account rotation to handle rate limits automatically.

dvcrn Updated Sep 8, 2026 32 repo stars

File contents

FlowForge

Autonomous spec → plan → code → QA pipeline powered by Claude Code. All heavy computation runs through Claude Code (Max subscription). OpenClaw only orchestrates.

Architecture

Flo (minimal tokens) → shell pipeline → Claude Code (all heavy work)
                                              ↓
                                   Account rotation on rate limit

Workflow Types

Classify the task before planning — each type has a different phase structure:

Type	When	Phase Order
`feature`	New capability	Backend → Worker → Frontend → Integration
`refactor`	Restructure existing code	Add New → Migrate → Remove Old → Cleanup
`investigation`	Bug hunt	Reproduce → Investigate → Fix → Harden
`migration`	Move data/infra	Prepare → Test → Execute → Cleanup
`simple`	Single-file change	Just subtasks, no phases

Steps

1. Setup workspace

bash ~/clawd/skills/flowforge/scripts/init_forge.sh "<task_description>" "<repo_path>"

Creates ~/.forge/<timestamp>/ with task.md.

2. Run the pipeline

bash ~/clawd/skills/flowforge/scripts/run_forge.sh ~/.forge/<timestamp>/

This chains 4 Claude Code calls:

Spec — generates spec.md (high thinking)
Plan — generates implementation_plan.json (high thinking)
Code — executes each subtask with verification (medium thinking)
QA — reviews output, scores against spec (high thinking)

Each step saves output to the workspace directory. Claude Code does ALL the work.

3. Monitor

Poll workspace for completion:

tail -f ~/.forge/<timestamp>/progress.log
cat ~/.forge/<timestamp>/qa_report.md

Account Rotation

Three Claude Max accounts rotate automatically on rate limit:

account-1@gmail.com  →  account-2@gmail.com  →  account-3@gmail.com  →  retry

Configure your accounts in ~/.flowforge/accounts.txt (one email per line). Save credentials per account in ~/.claude/accounts/<email>.json. Switch accounts with: bash <skill-dir>/scripts/rotate_account.sh

GitHub Issues

To pull a task from a GitHub issue:

gh issue view <number> --repo <owner>/<repo> --json title,body | \
  jq -r '"# " + .title + "\n\n" + .body' > ~/.forge/<timestamp>/task.md

Then run the pipeline normally.

Output

On completion, workspace contains:

spec.md — full specification
implementation_plan.json — phases + subtasks with status
qa_report.md — QA review and score
progress.log — timestamped execution log

Optional: Rubric Scoring (200 criteria)

Add --rubric flag for high-stakes runs. Scores against a universal 200-criterion quality rubric after the spec-based QA pass:

bash ~/clawd/skills/flowforge/scripts/run_forge.sh ~/.forge/<timestamp>/ --rubric

Rubric covers: Architecture (40), Code Quality (40), Testing (40), Error Handling (30), Security (20), Documentation (15), Observability (15).

Verdict thresholds: ≥180 = Ship it | 150–179 = Needs work | <150 = Major rework

Skip --rubric for quick tasks. Use it before shipping to production.

Prompts

See references/spec-prompt.md, references/planner-prompt.md, references/qa-prompt.md, references/rubric-prompt.md for the full Claude Code prompts used at each stage.

---name: flowforgedescription: Autonomous AI coding pipeline that breaks any task into a structured Spec → Plan → Code → QA loop and executes it via Claude Code. Use when: (1) starting any new feature, refactor, or bug fix, (2) given a GitHub issue to implement, (3) asked to "run FlowForge", "forge this", "plan and build", or "auto-implement". Routes ALL heavy work through Claude Code. Supports multi-account rotation to handle rate limits automatically.---# FlowForgeAutonomous spec → plan → code → QA pipeline powered by Claude Code.All heavy computation runs through Claude Code (Max subscription). OpenClaw only orchestrates.## Architecture```Flo (minimal tokens) → shell pipeline → Claude Code (all heavy work) ↓ Account rotation on rate limit```## Workflow TypesClassify the task before planning — each type has a different phase structure:| Type | When | Phase Order ||------|------|-------------|| `feature` | New capability | Backend → Worker → Frontend → Integration || `refactor` | Restructure existing code | Add New → Migrate → Remove Old → Cleanup || `investigation` | Bug hunt | Reproduce → Investigate → Fix → Harden || `migration` | Move data/infra | Prepare → Test → Execute → Cleanup || `simple` | Single-file change | Just subtasks, no phases |## Steps### 1. Setup workspace```bashbash ~/clawd/skills/flowforge/scripts/init_forge.sh "<task_description>" "<repo_path>"```Creates `~/.forge/<timestamp>/` with `task.md`.### 2. Run the pipeline```bashbash ~/clawd/skills/flowforge/scripts/run_forge.sh ~/.forge/<timestamp>/```This chains 4 Claude Code calls:1. **Spec** — generates `spec.md` (high thinking)2. **Plan** — generates `implementation_plan.json` (high thinking)3. **Code** — executes each subtask with verification (medium thinking)4. **QA** — reviews output, scores against spec (high thinking)Each step saves output to the workspace directory. Claude Code does ALL the work.### 3. MonitorPoll workspace for completion:```bashtail -f ~/.forge/<timestamp>/progress.logcat ~/.forge/<timestamp>/qa_report.md```## Account RotationThree Claude Max accounts rotate automatically on rate limit:```account-1@gmail.com → account-2@gmail.com → account-3@gmail.com → retry```Configure your accounts in `~/.flowforge/accounts.txt` (one email per line).Save credentials per account in `~/.claude/accounts/<email>.json`.Switch accounts with: `bash <skill-dir>/scripts/rotate_account.sh`## GitHub IssuesTo pull a task from a GitHub issue:```bashgh issue view <number> --repo <owner>/<repo> --json title,body | \ jq -r '"# " + .title + "\n\n" + .body' > ~/.forge/<timestamp>/task.md```Then run the pipeline normally.## OutputOn completion, workspace contains:- `spec.md` — full specification- `implementation_plan.json` — phases + subtasks with status- `qa_report.md` — QA review and score- `progress.log` — timestamped execution log## Optional: Rubric Scoring (200 criteria)Add `--rubric` flag for high-stakes runs. Scores against a universal 200-criterion quality rubric after the spec-based QA pass:```bashbash ~/clawd/skills/flowforge/scripts/run_forge.sh ~/.forge/<timestamp>/ --rubric```Rubric covers: Architecture (40), Code Quality (40), Testing (40), Error Handling (30), Security (20), Documentation (15), Observability (15).Verdict thresholds: **≥180 = Ship it** | **150–179 = Needs work** | **<150 = Major rework**Skip `--rubric` for quick tasks. Use it before shipping to production.## PromptsSee `references/spec-prompt.md`, `references/planner-prompt.md`, `references/qa-prompt.md`, `references/rubric-prompt.md` for the full Claude Code prompts used at each stage.

Frequently asked questions