
Security News
arXiv Is Rate Limiting Authors Following a Flood of AI Slop Submissions
arXiv now limits authors to two submissions a month as AI slop overwhelms moderators, delays good papers, and sparks debate over applying the limit to everyone.
software-factory
Advanced tools
Overnight Software Factory: Give it a GitHub issue, wake up to a deployed PoC. Powered by SuperPlane + Render.
Give it a GitHub issue or repo. Get a deployed PoC.
Your AI coding agent implements the issue. Software Factory deploys it to Render and opens the PR.
Built at SuperPlane Hackathon: Bash Script Funeral /w Render · NYC, June 27 2026
You: "Implement https://github.com/superplanehq/superplane/issues/5368"
Agent: fetch_github_issue → reads the issue
get_repo_structure → understands the codebase
✍️ writes the implementation (agent's own AI)
push_branch → pushes code to GitHub
deploy_preview → deploys to Render, returns live URL
create_pr → opens PR + comments preview URL on issue
Result: 🚀 Preview: https://factory-issue-5368.onrender.com
🔀 PR: https://github.com/superplanehq/superplane/pull/999
The AI agent you're already using does the code. Software Factory handles GitHub, Render, and SuperPlane — no Anthropic key required.
npm install -g software-factory
# or use directly:
npx software-factory
npx software-factory init
You'll be asked for 3 things:
| # | What | Where |
|---|---|---|
| 1 | SuperPlane API token | app.superplane.com → Profile → API Tokens |
| 2 | GitHub personal access token | github.com → Settings → Developer settings → PATs (repo scope) |
| 3 | Render API key | dashboard.render.com/u/settings → API Keys |
No Anthropic key needed. Your AI agent (Claude Code, Codex, OpenCode) is already the AI — software-factory is the infrastructure layer.
Or use environment variables for non-interactive setup:
export SUPERPLANE_TOKEN="TuovNZZl..."
export GITHUB_TOKEN="ghp_..."
export RENDER_API_KEY="rnd_..."
npx software-factory init --yes
# Register as MCP server (one time)
claude mcp add software-factory -- npx software-factory mcp
Then in any Claude Code session:
Use software-factory MCP tools to implement this issue and deploy it:
https://github.com/superplanehq/superplane/issues/5368
Steps:
1. factory_doctor — verify setup
2. fetch_github_issue — read the issue
3. get_repo_structure — understand the codebase
4. [implement the changes using your own tools]
5. push_branch — push your implementation
6. deploy_preview — get a live Render URL
7. create_pr — open PR + comment the preview URL on the issue
Add to your agent's MCP config (.mcp.json or agent settings):
{
"mcpServers": {
"software-factory": {
"command": "npx",
"args": ["software-factory", "mcp"]
}
}
}
Then ask your agent:
Use the software-factory tools to build and deploy a PoC for:
https://github.com/superplanehq/superplane/issues/5368
| Tool | Description |
|---|---|
factory_doctor | Check setup: SuperPlane, GitHub, Render all connected |
fetch_github_issue | Read issue title, body, labels, comments |
get_repo_structure | List files in a GitHub repo (any path, any branch) |
read_repo_file | Read a specific file from GitHub |
push_branch | Push code files to a new GitHub branch |
deploy_preview | Deploy branch to Render → returns live HTTPS URL |
get_deploy_status | Poll a Render deployment for live status |
create_pr | Open PR + comment preview URL on the original issue |
get_pipeline_status | Check SuperPlane canvas run history |
trigger_autonomous_pipeline | Run the full autonomous pipeline (no agent needed) |
sequenceDiagram
participant User
participant Agent as AI Agent<br/>(Claude Code / Codex)
participant MCP as Software Factory<br/>(MCP Tools)
participant SP as SuperPlane
participant GH as GitHub
participant RE as Render
User->>Agent: "Implement issue #5368 and deploy"
Agent->>MCP: fetch_github_issue(issue_url)
MCP->>GH: GET /repos/superplanehq/superplane/issues/5368
GH-->>Agent: Issue title, body, labels
Agent->>MCP: get_repo_structure(repo, path)
MCP->>GH: GET /repos/.../contents/
GH-->>Agent: File tree
Note over Agent: Agent reads relevant files,<br/>writes the implementation
Agent->>MCP: push_branch(repo, branch, files)
MCP->>GH: Create commit + branch
GH-->>Agent: Branch URL
Agent->>MCP: deploy_preview(repo, branch)
MCP->>RE: Create/update service, trigger deploy
RE-->>Agent: 🚀 https://xxx.onrender.com
Agent->>MCP: create_pr(repo, branch, preview_url)
MCP->>GH: POST /pulls + POST /issues/5368/comments
GH-->>Agent: PR URL
Agent->>User: Preview: https://xxx.onrender.com<br/>PR: https://github.com/.../pull/999
factory init [--yes] # One-time setup
factory doctor # Check configuration
factory build <url> [--follow] # Autonomous pipeline (no agent needed)
factory status [--watch] # Live pipeline status
factory logs # Per-stage execution logs
factory mcp # Start MCP server for AI agents
factory build --follow (Autonomous Mode)If you want the pipeline to run fully automatically without an AI agent session:
factory build https://github.com/superplanehq/superplane/issues/5368 --follow
This triggers the SuperPlane canvas which autonomously: fetches the issue → generates spec → writes code → deploys → opens PR. Requires an Anthropic API key stored in secrets (set during factory init).
flowchart LR
A([🚀 Start]) -->|issue_url\nrepo| B
B["📋 Fetch Issue\nGitHub API"]
B -->|passed| C
C["🧠 Requirement Agent\nGenerates spec"]
C -->|passed| D
D["⚙️ Implementation\nWrites + pushes code"]
D -->|passed| E
E["✅ Validation\nnpm test / build"]
E -->|passed| F
F["🚢 Deploy\nRender preview"]
F -->|passed| G
G["🔀 PR Agent\nPR + issue comment"]
style A fill:#7fe2b4,stroke:#333,color:#000
style G fill:#46E3B7,stroke:#333,color:#000
graph TB
subgraph agents["👤 AI Agents (already on your machine)"]
cc["Claude Code"]
cx["Codex"]
oc["OpenCode"]
any["Any MCP agent"]
end
subgraph factory["🏭 Software Factory (npm package)"]
mcp["MCP Server\nfactory mcp"]
cli["CLI\nfactory build"]
end
subgraph infra["☁️ Infrastructure"]
sp["SuperPlane\norchestration + state"]
gh["GitHub API\nfetch · push · PR"]
re["Render API\ndeploy · preview URL"]
end
cc & cx & oc & any -->|"MCP tools"| mcp
cli -->|"trigger_autonomous"| sp
mcp --> gh
mcp --> re
mcp --> sp
sp --> gh
sp --> re
style agents fill:#1a1a2e,color:#eee,stroke:#7fe2b4
style factory fill:#16213e,color:#eee,stroke:#46E3B7
style infra fill:#0f3460,color:#eee,stroke:#7fe2b4
The factory was built to solve these 5 SuperPlane open source issues:
# In your AI agent:
# "Use software-factory MCP tools to implement and deploy:"
factory build https://github.com/superplanehq/superplane/issues/5368 --follow # Markdown + Mermaid
factory build https://github.com/superplanehq/superplane/issues/5366 --follow # Canvas version diff
factory build https://github.com/superplanehq/superplane/issues/5164 --follow # Send execution to chat
factory build https://github.com/superplanehq/superplane/issues/5704 --follow # Run inspection UX
factory build https://github.com/superplanehq/superplane/issues/5705 --follow # Canvas warnings
graph LR
subgraph bin["bin/"]
fjs["factory.js\nCLI entrypoint"]
end
subgraph src["src/"]
subgraph mcp["mcp/"]
ms["server.js\n10 MCP tools\nfetch · push · deploy · PR"]
end
subgraph commands["commands/"]
ci["init.js\nsetup wizard"]
cd["doctor.js\nhealth checks"]
cb["build.js\nautonomous trigger + follow"]
cs["status.js\nlive status"]
cl["logs.js\nexecution logs"]
end
subgraph sp["superplane/"]
client["client.js\nREST client"]
tmpl["canvas-template.js\n7-node pipeline"]
end
cfg["config.js\n~/.factory/config.json"]
end
fjs --> ci & cd & cb & cs & cl & ms
ms --> client & cfg
ci --> tmpl & client & cfg
style bin fill:#2d2d2d,color:#eee,stroke:#7fe2b4
style src fill:#1a1a2e,color:#eee,stroke:#46E3B7
git clone https://github.com/hongchengw/superplane-render-nyc-hack
cd superplane-render-nyc-hack
npm install
node bin/factory.js --help
MIT © Roshan Sharma
Built with SuperPlane · Deployed on Render
FAQs
Overnight Software Factory: Give it a GitHub issue, wake up to a deployed PoC. Powered by SuperPlane + Render.
The npm package software-factory receives a total of 20 weekly downloads. As such, software-factory popularity was classified as not popular.
We found that software-factory demonstrated a healthy version release cadence and project activity because the last version was released less than a year ago. It has 1 open source maintainer collaborating on the project.

Security News
arXiv now limits authors to two submissions a month as AI slop overwhelms moderators, delays good papers, and sparks debate over applying the limit to everyone.

Research
/Security News
A new GhostAction wave hits hundreds of GitHub repos, expanding CI/CD secret theft to cloud and AI credentials in source code and git history.

Research
/Security News
Tensorlake npm SDK version 0.5.144 was compromised in a ChainDrop / Shai-Hulud attack, delivering credential-stealing malware.