aronprins

aronprins/paperclip-company-playbook

Skill Claude CodeCodex

A mandatory skill for closing the loop after every Founder correction. Ensures mistakes are logged, instructions updated, and fixes verified.

aronprins/paperclip-company-playbook

Skill Claude CodeCodex

A mandatory verification skill for CEO agents. Ensures every deliverable is personally verified before reporting to the Founder.

aronprins/paperclip-company-playbook

Skill Claude CodeCodex

A mandatory verification skill for worker agents. Ensures every agent checks their own work before marking a task as done.

aronprins/claude-loop

Skill Claude CodeCodex

Autonomous orchestrated loop that works through a prd.json story-by-story using a fresh Task subagent per story. Use this skill whenever the user asks to "run the loop", "claude loop", "start the loop", "kick off the loop", "work through the prd", "process prd.json", "implement all the stories", "iterate through the…

aronprins/codex-loop

Instructions file CodexOpenCode

Instructions for aronprins/codex-loop, a project described as: Hand Codex a PRD. Get a working feature back. One story per subagent, fresh context every time.

aronprins/codex-loop

Skill Claude CodeCodex

Autonomous Codex workflow that implements codex-loop/prd.json with one fresh Codex subagent per story. Use when the user asks to run Codex Loop, run the loop, work through a PRD, process prd.json, implement all stories, iterate through user stories autonomously, or complete a feature broken into PRD stories.…

aronprins/paperclip-cli

Skill Claude CodeCodex

Operates the Paperclip CLI (paperclipai) to run a Paperclip instance and manage an AI-agent company from the terminal — companies, agents, issues, goals, projects, runs, routines, approvals, budgets, secrets, and plugins. Use whenever the user says "use the Paperclip CLI", "paperclipai", asks to bootstrap or operate a…

aronprins/run-paperclip

Plugin Claude Code

Run your Paperclip company from Claude Code — operate the board and embody agents on your own subscription.

aronprins/run-paperclip

Plugin Claude Code

Run your Paperclip company from Claude Code — operate the board and embody agents on your own subscription, instead of server-side claude -p.

aronprins/run-paperclip

Instructions file CodexOpenCode

AGENTS.md instructions for aronprins/run-paperclip, covering run-paperclip — conventions for every subagent, scope & boundaries, the one rule that must never break (cost-safety), secrets boundary (hard constraint) and concurrency & orchestration.

aronprins/run-paperclip

Agent

Embodies a Paperclip engineering agent to execute ONE assigned issue locally — checkout, do the engineering work, post results, set disposition, release, record cost. Dispatched by /pc-run-company.

aronprins/run-paperclip

Agent

Embodies a Paperclip manager agent to coordinate ONE assigned issue locally — triage, break it down into child issues, assign, set disposition, release, record cost. Does NOT execute engineering. Dispatched by /pc-run-company.

aronprins/run-paperclip

Agent

Embodies a Paperclip reviewer agent to review ONE assigned issue's work products / PR locally — checkout, review, post review comments, set disposition (approve=done / request-changes=inprogress), release, record cost. Dispatched by /pc-run-company.

aronprins/run-paperclip

Command

Make the calls only you can make — review what your agents are asking for (new hires, strategy changes) and approve, reject, ask for changes, or comment. The one place you sign off on decisions.

aronprins/run-paperclip

Command

Make a company cost-safe so its work runs on your Claude subscription, never billed to you separately. Use --undo to turn it back off.

aronprins/run-paperclip

Command

Keep all your companies moving at once — scan every company for ready work and do it on your Claude subscription, so you can stay in the Paperclip UI while this picks up and finishes the work. Take a quick look (snapshot), do one pass (run), or keep it running (start/stop/status).

aronprins/run-paperclip

Command

Do a whole company's ready work in this session, on your subscription — you step into each owning agent and finish its issues. Add --loop to keep the open session picking up new work hands-off.

aronprins/run-paperclip

Command

Keep one company running hands-off in THIS session — leave it open and it picks up and finishes new work on its own, on your subscription, until you close it. start | stop | status.

aronprins/run-paperclip

Command

Get set up to run your company in Paperclip — connect, choose how you'll use it, and confirm everything works. Run this once to start.

aronprins/run-paperclip

Command

Start here — pick what you want and I'll ask only what I need.

aronprins/run-paperclip

Command

Work through your agent's assigned issues yourself, one at a time, posting results and recording cost as you go. Runs on your Claude subscription.

aronprins/run-paperclip

Hook

Runs before the agent uses a tool for Bash tool calls, executing cost-guard.sh and destructive-guard.sh (2 commands). From aronprins/run-paperclip.

aronprins/run-paperclip

Skill Claude CodeCodex

Act AS a Paperclip agent and execute its assigned work locally: read the inbox or a wake payload, checkout the issue, load context, do the real work in this session, post results, set the final disposition, release, and record a cost event. Use when this session is embodying a Paperclip agent (PAPERCLIPAGENTID is set)…

aronprins/run-paperclip

Skill Claude CodeCodex

How to operate a Paperclip company from Claude Code: the board-operator vs embodied-agent personas, the cost-safe rule (never wake a claudelocal agent server-side — embody locally instead), and the CLI/API command surface. Use whenever the user asks to run, operate, drive, or embody a Paperclip company, hire agents…