Getting it into your agent
One page per mod, every tool's command on it. A separate URL per tool would split the same page into five that compete with each other.
npx agentmods add commands/third-order-labs/longform-plugin/start-projectgit clone --depth 1 https://github.com/third-order-labs/longform-pluginWhat 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.
| Model | Per session | Once invoked |
|---|---|---|
| Fable 5 | $0.00026 | $0.02008 |
| Opus 5 | $0.00013 | $0.01004 |
| Sonnet 5 | $0.00005 | $0.00402 |
| Haiku 4.5 | $0.00003 | $0.00201 |
Grade A, and why
start-project scanned grade A with 0 findings 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 yesterday.
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.
Nothing flagged
None of the 26 patterns this scan looks for appear in this file: no shell pipes, no recursive deletes, no credential paths, no hidden text, no instruction-override or anti-refusal phrasing, no agent-config snooping. That is not a guarantee, it is the absence of the things that are checkable.
How it starts
The opening of the file, as written. The whole thing — 146 lines — stays where its author put it; the contents beside it link to each section on GitHub.
/start-project — Start a New Writing Project
Start a new project through conversation. Figure out the story with the user, capture enough to start writing, and get out of the way. The goal is not to build infrastructure — it's to reach a state where /plan-chapter makes sense for chapter one.
Documents and tracking systems emerge later as the project needs them. Don't front-load structure the user hasn't earned a reason to want yet.
Invocation
/start-project
Skills Referenced
project-scaffolding— directory conventions and document templates (used selectively, not exhaustively)long-form-methodology— workflow philosophy and document hierarchyauthorial-voice— voice capture approach (if user wants to establish voice upfront)
Workflow
Phase 1: Have the Conversation
This is the only phase that matters. Everything else is just writing down what you figured out here.
Start by letting the user talk about their project. They might show up with a detailed outline or a half-formed feeling. Meet them where they are.
What you need to understand before moving on:
- What's the story about? (not a synopsis — what's it about)
- Who's the main character and what's their problem?
- What kind of book is this? (genre, tone, scope — standalone or series)
- What's the world like? (broad strokes only — enough to set chapter one somewhere)
What you don't need yet:
- A complete cast of characters (they'll appear as you write)
- Full world-building (it fills in through chapters)
- Locked POV/tense/style decisions (these can emerge from writing the first few chapters — some users know, some discover)
- An act structure or beat sheet (some users want this, some want to find the shape as they go)
How to run this conversation:
- Let the user riff. Ask follow-up questions, not checklists.
- If they're detailed and prepared, move fast — don't slow them down with questions they've already answered.
- If they're vague, help them develop the idea. Push on the protagonist's problem and the core tension — that's usually where a vague concept becomes a real story.
- If they start going deep on world-building or backstory, let them — but recognize when you have enough to start. You can always build more world later. You can't write chapter one without a character and a situation.
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.
- yesterday First seen · 146 lines · 26 tokens per session scan A 427a79768b21
start-project is a command published in the GitHub repository third-order-labs/longform-plugin (2 stars, last pushed 1mo ago), licensed MIT. It adds 26 tokens to every session and 2,008 once invoked, about $0.0001 per session on Opus 5. A static security scan graded it A with 0 findings. No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-31.
Other commands, from other repositories
onboard-interactive
Interactive customer onboarding with Step 0 independent verification (dialog-based, 17 stagegates).
monitor
AML transaction monitoring and suspicious activity reporting.
skip_sample
Skip sample-first mode for this request — fetch/enrich at full scale without a 5-entity preview.
dev-flow
从 PRD 到交付的完整 8 步研发流程.
axonflow-status
Show your AxonFlow tenantid (needed for Stripe Pro checkout), endpoint, tier (Free vs Pro), and Pro license expiry date.
unit-economics
Unit economics analysis — CAC/LTV estimation, churn inference, gross margin per unit.