openproblems

Project instructions for OpenProblems, a benchmark project for comparing methods that analyse spatial transcriptomics data—data showing where gene activity occurs inside tissue. It covers an analysis server and a system for sharing instructions across coding agents.

In plain words
What is it for?
Use it when working on the OpenProblems server, its bioinformatics validation and inspection tools, or the installer that adapts project instructions for different coding agents.
Why use it?
It gives coding agents one source of rules and clarifies which analysis tools exist, reducing inconsistent instructions and preventing them from trying to run unsupported pipelines.

Cursor rule for Cursor

Install

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.

agentmods
npx agentmods add rules/rebell-leader/spatialai_mcp/openproblems
Clone the repo
git clone --depth 1 https://github.com/Rebell-Leader/SpatialAI_MCP

Made for: Cursor.

Per session 1,429 This file is loaded in full into every session.
When invoked 1,429 The same file — it is already loaded in full.
Security scan A 0 findings. Scan, not verified.
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 $0.01429 $0.01429
Opus 5 $0.00714 $0.00714
Sonnet 5 $0.00286 $0.00286
Haiku 4.5 $0.00143 $0.00143

Measured 2d ago against content hash 1d8488dd662c, method: parsed. Prices are Anthropic first-party input rates as of 2026-08-30, from the pricing page.

Security

Grade A, and why

openproblems 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.

.cursor/rules/openproblems.mdc · 113 lines

How it starts

The opening of the file, as written. The whole thing — 113 lines — stays where its author put it; the contents beside it link to each section on GitHub.

AGENTS.md — OpenProblems Spatial Transcriptomics Co-pilot

Canonical, provider-neutral instructions for any AI coding agent working in this repository (Claude Code, OpenAI Codex, Cursor, GitHub Copilot, and others). This is the single source of truth. Tool-specific files (CLAUDE.md, .cursor/rules/, .github/copilot-instructions.md) are generated from this file and the skills/ directory by installer/ — do not edit them by hand.

What this project is

A co-pilot for computational biologists working on the OpenProblems spatial transcriptomics benchmarks. It ships two things:

  1. An MCP server (src/openproblems_mcp/) — domain-aware validation and analysis tools for spatial data and bioinformatics workflows, exposed over the Model Context Protocol so any MCP-capable agent can call them.
  2. A provider-agnostic skill installer (skills/, context/, installer/) — author rules/skills/context once here, then project them into whichever agent the user runs.

Ground truth: which MCP tools actually exist

Only call tools that are registered. As of now the server is read-only analysis — it validates and inspects, it does not execute pipelines.

Available now: health_check, list_tools_status, get_server_info, validate_spatial_data, validate_multiple_spatial_files, analyze_spatial_metadata, check_spatial_data_compatibility, extract_bioinformatics_metadata, analyze_workflow_configuration, assess_data_quality, analyze_workflow_dependencies.

Roadmap — NOT implemented, do not call: run_nextflow_workflow, run_viash_component, build_docker_image, run_nf_test, analyze_nextflow_log, create_spatial_component, setup_spatial_environment, and the OpenProblems build/benchmark/submission tools. For these actions, drive the user's local nextflow / viash / docker CLIs directly via the terminal instead, and say so plainly.

Read the full file on GitHub · 113 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. 2d ago First seen · 113 lines · 1,429 tokens per session scan A 1d8488dd662c

Subscribe to this mod's changes

openproblems is a cursor rule published in the GitHub repository Rebell-Leader/SpatialAI_MCP (4 stars, last pushed 2mo ago), licensed Apache-2.0. It adds 1,429 tokens to every session, about $0.0071 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.