inkwell-agent-sandboxes-and-software-factory: Command for Claude Code

.claude/commands/prime.md

prime is a command for Claude Code from disler/inkwell-agent-sandboxes-and-software-factory. It costs 33 tokens per session (1,080 once invoked), scanned A, original, MIT.

A command that gives an agent the background needed to work on Inkwell, its software-building system, and the disposable virtual machines that run it.

In plain words
What is it for?
Use it at the start of work on this system to inspect its main layers, commands, files, and operating model.
Why use it?
It prevents the agent from starting without understanding how the application, workflow engine, and virtual-machine setup fit together.

Command for Claude Code

Written for Claude Code: installed under .claude/. Also seen: reads .claude/ paths; positional $N argument.

This is disler/inkwell-agent-sandboxes-and-software-factory's own configuration. It tells Claude Code how to work on inkwell-agent-sandboxes-and-software-factory itself, so it is not a mod to install elsewhere. Copy it as a starting point and replace the rules that are about this project. Everything inkwell-agent-sandboxes-and-software-factory configures →

Reuse

Borrowing it

Nothing to install: this file belongs to disler/inkwell-agent-sandboxes-and-software-factory. Take a copy, put it at the same path in your own repository, and replace the rules that are about this project with yours.

Copy the file
curl -O https://raw.githubusercontent.com/disler/inkwell-agent-sandboxes-and-software-factory/main/.claude/commands/prime.md
Clone the repo
git clone --depth 1 https://github.com/disler/inkwell-agent-sandboxes-and-software-factory

Made for: Claude Code.

Wrote 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.

agentmods badge for prime

README.md
[![agentmods](https://agentmods.dev/badge/commands/disler/inkwell-agent-sandboxes-and-software-factory/prime/github.svg)](https://agentmods.dev/commands/disler/inkwell-agent-sandboxes-and-software-factory/prime)
Your own site
<a href="https://agentmods.dev/commands/disler/inkwell-agent-sandboxes-and-software-factory/prime"><img src="https://agentmods.dev/badge/commands/disler/inkwell-agent-sandboxes-and-software-factory/prime/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.

agentmods 80×15 button for prime

Your own site · 80×15
<a href="https://agentmods.dev/commands/disler/inkwell-agent-sandboxes-and-software-factory/prime"><img src="https://agentmods.dev/badge/commands/disler/inkwell-agent-sandboxes-and-software-factory/prime.svg" alt="Reviewed on agentmods" width="80" height="20"></a>
Per session 33 Only the description is in the session, so the agent can decide to use it. The body loads when it is invoked.
When invoked 1,080 The whole file, excluding the scripts and references it only reads on demand.
Security scan A 1 finding. A grade says what 26 rules found in the file — not that it is safe.
Origin original No closer match found in the catalogue.
Token cost

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.

ModelPer sessionOnce invoked
Fable 5.1 $0.00033 $0.01080
Opus 5 $0.00016 $0.00540
Sonnet 5 $0.00007 $0.00216
Haiku 4.5 $0.00003 $0.00108

Measured 13d ago against content hash 669a3839ac05, method: parsed. Prices are Anthropic first-party input rates as of 2026-09-12, from the pricing page.

Security

Grade A, and why

prime scanned grade A with 1 finding 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 13d 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.

Makes network callslowCapability

Not a fault in itself. Listed so you know the mod talks to something, and to what.

6. Read `.claude/skills/sssf-sandbox-orchestrator/SKILL.md` and its `references/gotchas.md`. The governing rule is **thin skill, fat recipes**: every action should be a `just` command a human could type. Dropping to `ssh
.claude/commands/prime.md · 30 lines

How it starts

The opening of the file, as written. The whole thing — 30 lines — stays where its author put it; the contents beside it link to each section on GitHub.

Purpose

Orient yourself in a three-layer system: Inkwell (a small blog-writing app), the Super Simple Software Factory (deterministic Python owns the graph, coding agents are bounded phases inside it), and the sandbox mount system (six host-side phases that stand the whole thing up on a disposable exe.dev VM, run it, and watch it from outside). The app is the payload; the point is the loop that ships it without a human in the middle.

Workflow

  1. Map the surface first, because it is the fastest way to see the shape: just (four namespaces, nothing else), then just --list sbx, just --list adw, just --list obs, just --list local. The namespace answers where the work happens: sbx orchestrates VMs from the host, adw runs the workflows, local boots an orchestrator agent on this machine, obs reads the trace db. Then git ls-files | head -60 and ls sandbox_mount/host sandbox_mount/guest just/sandbox.

  2. Read TREE.md — every file that matters and why it exists, grouped by layer, ending with the five things that will bite you. It is the map; the rest of this workflow is the territory. Then README.md for the three layers, then specs/sandbox-mount-system.htmlthe plan is the working checklist, with live checkboxes recording what has actually been verified on hardware versus what is still theory. Open it in a browser or read the data-ck="... checked" attributes. Treat an unchecked box as "not proven", not "not written".

  3. Read ai_docs/exedev_sandbox_mounting.md — every number in it was measured on live VMs, and several "obvious" designs were killed by those measurements (a custom Docker image buys ~1s; --setup-script cannot see --env; an unsynced ssh exe.dev cp silently produced 5,641 zero-byte files). Do not re-derive these; do not contradict them without a new measurement.

  4. Read justfile and just/sandbox/mod.just. This is the load-bearing idea: just/sandbox/ and sandbox_mount/host/ are host-only — not because they are removed from the mounted copy (the whole repo ships intact) but because they need the exe.dev account and OPENROUTER_PROVISIONING_KEY, and neither credential ever leaves the host. A sandbox that runs just sbx mount gets an auth failure, not a nested VM.

Read the full file on GitHub · 30 lines

Changes

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.

  1. 13d ago First seen · 30 lines · 33 tokens per session scan A 669a3839ac05

Subscribe to this mod's changes

prime is a command published in the GitHub repository disler/inkwell-agent-sandboxes-and-software-factory (154 stars, last pushed 1mo ago), licensed MIT. It adds 33 tokens to every session and 1,080 once invoked, about $0.0002 per session on Opus 5. A static security scan graded it A with 1 finding (makes network calls). No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-30.