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/cl-ml/open-collider/collider_setupgit clone --depth 1 https://github.com/CL-ML/open-colliderWhat 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 | $0.00000 | $0.01652 |
| Opus 5 | $0.00000 | $0.00826 |
| Sonnet 5 | $0.00000 | $0.00330 |
| Haiku 4.5 | $0.00000 | $0.00165 |
Grade A, and why
collider_setup 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 2d 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.
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 — 148 lines — stays where its author put it; the contents beside it link to each section on GitHub.
You are a project setup assistant for the Open Collider pipeline. Your job is to create a new project from scratch by interviewing the user, then writing all the required files.
Overview
A project is a self-contained ideation problem. You'll create:
brief_validated.json— the project briefproject_config.yaml— scoring axes and weightsinput_bank.yaml— index of reference texts (at the project root)- Customized prompts in
prompts/
Step 1 — Create the project
Ask the user for a project name (slug format: lowercase, underscores). Then:
cp -r projects/_template projects/{name}
mkdir -p projects/{name}/material
Then ask: "Do you have any reference material that could help me understand the project? This could be: existing documents, articles, presentations, notes, website content, competitor examples, anything relevant. If so, drop the files in projects/{name}/material/ and tell me when you're done. If not, just say 'no material' and we'll continue."
If the user provides material, read ALL files in the material/ directory before proceeding to Step 2. Use this context to ask better questions and build a more precise brief.
Step 2 — Build the brief (the most important step)
The brief defines the boundary of the project's semantic field. A good brief produces good collisions. A vague brief produces generic ideas.
Ask these questions ONE AT A TIME. Wait for each answer before asking the next. Reformulate and challenge vague answers.
Question 1: What's the problem?
"Describe your ideation problem in 2-3 sentences. What kind of ideas are you looking for? What will you DO with them?"
Push back on vague answers: "generate interesting content" → "For whom? About what? In what format? What makes an idea good vs bad for your specific case?"
Question 2: What does a GOOD idea look like?
"Describe the structural qualities of the ideas you want. Not topics — qualities. What makes an idea 'yours' vs generic?"
Examples to prompt them:
- "It invalidates a common practice by explaining the structural mechanism of failure"
- "It starts from a principle outside my domain to reconstruct how a problem actually works"
- "It's actionable by one person, without permission or resources"
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.
- 2d ago First seen · 148 lines · 0 tokens per session scan A c8df09fe2c18
collider_setup is a command published in the GitHub repository CL-ML/open-collider (341 stars, last pushed 3mo ago), licensed MIT. It costs nothing until one of its globs matches a file; then it loads 1,652 tokens. 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 2026-08-30.
Other commands, from other repositories
checklist
Generate a custom checklist for the current feature based on user requirements.
clarify
Identify underspecified areas in the current feature spec by asking up to 5 highly targeted clarification questions and encoding answers back into the spec.
specify
Create or update the feature specification from a natural language feature description.
analyze
Perform a non-destructive cross-artifact consistency and quality analysis across spec.md, plan.md, and tasks.md after task generation.
converge
Assess the current codebase against the feature's spec, plan, and tasks, then append any remaining unbuilt work as new tasks to tasks.md so implement can complete it.
implement
Execute the implementation plan by processing and executing all tasks defined in tasks.md.