DEV Community: DasClown

eu-regulation-mcp: Track EU laws, rulings & national gazettes via MCP

DasClown — Wed, 13 May 2026 21:03:47 +0000

eu-regulation-mcp — EU Regulation Intelligence MCP Server

https://github.com/DasClown/eu-regulation-mcp

An MCP server that monitors EU legislation across multiple sources: EUR-Lex, EU Parliament, European Commission, ECJ, and national gazettes.

What it monitors

EUR-Lex — EU treaties, directives, regulations, decisions
EU Parliament — Legislative proposals, amendments, resolutions
European Commission — Delegated acts, implementing acts, consultations
ECJ — Court rulings and opinions
National gazettes — Member state implementations

Who is it for

Legal teams tracking regulatory changes
Compliance officers monitoring sector-specific legislation
Policy analysts following EU lawmaking

MIT licensed, open-source.

drug-pipeline-mcp: Open-source pharmaceutical R&D intelligence

DasClown — Wed, 13 May 2026 21:03:36 +0000

drug-pipeline-mcp — Pharmaceutical R&D Pipeline Intelligence

https://github.com/DasClown/drug-pipeline-mcp

An MCP server that gives AI agents full visibility into pharmaceutical R&D.

What it covers

Clinical trials — Search ClinicalTrials.gov by condition, phase, sponsor, or status
FDA/EMA approvals — Full approval history and EU authorization status
Safety data — FAERS adverse event reports with PRR signal detection
Drug labels — Full prescribing information, interactions, contraindications
Patent expiry — Orange Book patent data and exclusivity timelines
Recalls — Active and past FDA recalls with classification
Publications — PubMed search for clinical evidence

How to use

Add it to your MCP config and ask your AI: "What's in Novo Nordisk's pipeline?"

MIT licensed, open-source.

climate-csrd-mcp: Open-source CSRD climate compliance for AI agents

DasClown — Wed, 13 May 2026 20:57:51 +0000

climate-csrd-mcp — EU CSRD Climate Intelligence MCP Server

https://github.com/DasClown/climate-csrd-mcp

An MCP server purpose-built for EU CSRD (Corporate Sustainability Reporting Directive) compliance. Climate data, ESG metrics, and regulatory reporting — all accessible to AI agents.

Why this matters

The EU CSRD requires thousands of companies to report on climate impact, ESG metrics, and sustainability risks. This MCP server makes that data machine-accessible.

Key capabilities

Climate data — Temperature records, emission factors, climate risk zones
ESG metrics — Environmental, social, and governance indicators
Regulatory context — CSRD reporting standards and deadlines
Sector benchmarks — Compare against industry baselines

Integration

Works as a standard MCP server. Connect via:

mcpServers:
  climate-csrd:
    command: uvx
    args: [climate-csrd-mcp]

Who needs it

Sustainability officers preparing CSRD reports
Auditors verifying ESG disclosures
Investors doing climate due diligence
Consultants advising on EU compliance

Open-source, MIT licensed.

CropProphEU: An open-source MCP server for EU crop intelligence

DasClown — Wed, 13 May 2026 20:57:44 +0000

CropProphEU — EU Crop Intelligence MCP Server

https://github.com/DasClown/CropProphEU

An MCP server providing comprehensive crop intelligence for EU agriculture. Designed for AI agents that need real-time agricultural data.

What it does

Yield forecasts — Predicts crop yields based on soil data, climate models, and historical patterns
Market values — Current market prices and trends for EU agricultural commodities
Risk analysis — Assesses climate risks, pest pressures, and supply chain vulnerabilities

How to use

# Add to your MCP config
{
  "mcpServers": {
    "crop-mcp": {
      "command": "uvx",
      "args": ["crop-proph-eu"]
    }
  }
}

Then ask your AI agent: "What's the wheat yield forecast for northern France?"

Use cases

Insurance companies assessing crop risk
Agricultural traders monitoring EU markets
Farmers planning seasonal strategies
Policy analysts tracking food security

Open-source, MIT licensed, and available via Smithery and Glama.

5 open-source MCP servers I built for EU agriculture, pharma, climate & law

DasClown — Wed, 13 May 2026 20:47:52 +0000

I have been building Model Context Protocol (MCP) servers for EU domains. Here is the full collection.

1. CropProphEU — EU Crop Intelligence

https://github.com/DasClown/CropProphEU

An MCP server for crop yield forecasts, market values, and risk analysis across EU agriculture. Integrates soil data, climate models, and market prices.

2. drug-pipeline-mcp — Pharmaceutical R&D Pipeline

https://github.com/DasClown/drug-pipeline-mcp

Pharmaceutical R&D Pipeline Intelligence MCP Server. Covers clinical trials, FDA/EMA approvals, safety data (FAERS), drug interactions, patent expiry, and publications.

3. climate-csrd-mcp — CSRD Climate Compliance

https://github.com/DasClown/climate-csrd-mcp

MCP server for EU CSRD (Corporate Sustainability Reporting Directive) compliance — climate data, ESG metrics, and regulatory reporting.

4. eu-regulation-mcp — EU Regulation Tracker

https://github.com/DasClown/eu-regulation-mcp

EU Regulation Intelligence MCP Server. Monitors EUR-Lex, EU Parliament, Commission, ECJ, and national gazettes.

5. awesome-mcp-servers — Curated List

https://github.com/DasClown/awesome-mcp-servers

A curated collection of MCP servers spanning agriculture, pharma, climate, regulation, and dev tools.

All servers work as standalone MCP processes or via Smithery/Glama.

I audited 10 Pre-Mortem tools — then built one that actually works

DasClown — Wed, 13 May 2026 20:31:11 +0000

The 5-minute exercise that saved me 6 months

Last week I had a business idea. D2C flower dropshipping. 20 tulips, fast shipping, great margins on paper. I was ready to build.

A friend said: "Premortem it first."

I spent 5 minutes imagining it already failed 6 months from now. The results:

❌ Cold chain logistics kills the margin

❌ VAT + EORI across EU countries eats the rest

❌ Return rate on perishables is uninsurable

❌ Even established players can't make online flowers work

Verdict: Don't build.

5 minutes vs 6 months of learned lessons. That's the value of a Pre-Mortem.

What's a Pre-Mortem?

Instead of asking "will this work?" → imagine it already failed and explain why.

Gary Klein published it in Harvard Business Review (2007). Daniel Kahneman called it "my single most valuable decision technique." The magic is in prospective hindsight — when you tell someone "this already failed, explain why," their brain generates 30% more specific failure reasons than when you ask "what could go wrong."

For AI-assisted decisions, this matters even more. LLMs default to agreeable, optimistic responses. Ask "is this a good plan?" and it finds reasons to say yes. A Pre-Mortem flips the frame.

The Audit: I checked 10 repos. Best was 2/4.

I wanted an AI Pre-Mortem skill that actually worked. So I audited every repo I could find on GitHub, HuggingFace, GitLab, and npm.

The 4 essential features:

#	Feature	Why it matters
1	Base Rates	Real failure statistics, not just "think about base rates"
2	Bias Circuit-Breaker	Systematic checks for Sycophancy, Optimism, Availability, Anchoring
3	L/I Scoring	Quantitative Likelihood × Impact (1-5), not qualitative hand-waving
4	Commitment	A concrete action with a date, not "consider X"

Results

Repo	Base Rates	Bias CB	L/I Score	Commitment
Hi1talib1World/Premortem ⭐51	❌	❌	❌	❌
AndyShaman/premortem ⭐16	❌	⚠️	❌	❌
MADEVAL/Pre-Mortem-Skill ⭐2	❌	❌	✅	✅
MrBinnacle/azimuth ⭐5	✅	⚠️	❌	❌
b1rdmania/claude-premortem-skill ⭐1	❌	❌	❌	❌
premortem-skill (this one)	✅	✅	✅	✅

None had more than 2/4. The field was wide open.

What I built: `premortem-skill`

GitHub: DasClown/premortem-skill

A Claude Code skill with two modes:

Quick Mode (`!pm`) — 30 seconds

!pm
→ 3 questions + quick L/I calibration
→ Output: ~10 lines
→ For day-to-day decisions during coding

Full Mode (`!pm full`) — 2 minutes

!pm full
→ All 4 features
→ Risk Matrix with L×I scores
→ Bias Circuit-Breaker report
→ Concrete commitment (action + date)

How the 4 features work together

1. Base Rates 🎲

Real statistics, not generic advice. When you say "this refactor takes 2 weeks," the skill checks:

Decision Class	Failure Rate	Source
Software timeline overshoot	70%	Standish CHAOS
Refactoring exceeds 50%+ estimate	64%	IEEE
Product launch failure	70-95%	Nielsen
Feature without user research	64% fail	Pragmatic Institute

The 2-week estimate gets recalibrated to 3-4 weeks. That's not pessimism — that's reality.

2. Bias Circuit-Breaker 🛑

After generating failure modes, 4 systematic checks run automatically:

Bias	Question	Correction
Sycophancy	"Would I say this to a stranger?"	Remove politeness filter
Optimism	"Inside vs. outside view?"	Base rate recalibration
Availability	"Is this from my last project?"	Demand evidence
Anchoring	"Where does this number come from?"	Independent re-estimate

3. L/I Scoring 📊

Every failure mode gets a Likelihood × Impact score:

Score	Zone	Action
1-4	🟢 Green	Accept
5-8	🟡 Yellow	Monitor
9-12	🟠 Orange	Needs mitigation
15-25	🔴 Red	STOP. Rethink the plan.

4. Commitment 🎯

No "consider X" allowed. Every full Pre-Mortem ends with one concrete action by a date.

✅ "Write integration tests for payment flow by Thursday 18:00"

❌ "Consider testing more"

Real example

When I Pre-Mortemed a planned auth module refactor:

🔪 PREMORTEM
Plan: Refactoring auth from JWT to session tokens

1. Most Likely → Timeline undercounted
   Base Rate: 64% of refactors exceed 50%+
   L=4 I=4 → 16 🔴

2. Worst-Case → Session invalidation breaks active users
   L=2 I=5 → 10 🟠

3. Bias Check → Optimism detected (inside: 2w, outside: 3-4w)

4. Commitment → Write migration script + dry-run on staging first
   → By: Thursday

The base rate forced me to double my estimate. The commitment made me write the migration script before touching anything. The whole thing took 2 minutes.

How to use it

# Install (30 seconds)
git clone https://github.com/DasClown/premortem-skill.git ~/.claude/skills/premortem

# Use in Claude Code
!pm              # Quick: 30 seconds
!pm full         # Full: 2 minutes, all 4 features

Why this matters

Every developer knows the feeling: "I should have seen that coming."

Pre-Mortems are the antidote. Not because you're pessimistic — because you're honest before reality forces honesty on you.

30 seconds now saves hours later.

GitHub: DasClown/premortem-skill

Full audit data: AUDIT.md

Base rates reference: references/base-rates.md

Built with Kahneman's prospective hindsight, Klein's HBR method, and Tetlock's superforecasting principles.

DEV Community: DasClown

eu-regulation-mcp: Track EU laws, rulings & national gazettes via MCP

eu-regulation-mcp — EU Regulation Intelligence MCP Server

What it monitors

Who is it for

drug-pipeline-mcp: Open-source pharmaceutical R&D intelligence

drug-pipeline-mcp — Pharmaceutical R&D Pipeline Intelligence

What it covers

How to use

climate-csrd-mcp: Open-source CSRD climate compliance for AI agents

climate-csrd-mcp — EU CSRD Climate Intelligence MCP Server

Why this matters

Key capabilities

Integration

Who needs it

CropProphEU: An open-source MCP server for EU crop intelligence

CropProphEU — EU Crop Intelligence MCP Server

What it does

How to use

Use cases

5 open-source MCP servers I built for EU agriculture, pharma, climate & law

1. CropProphEU — EU Crop Intelligence

2. drug-pipeline-mcp — Pharmaceutical R&D Pipeline

3. climate-csrd-mcp — CSRD Climate Compliance

4. eu-regulation-mcp — EU Regulation Tracker

5. awesome-mcp-servers — Curated List

I audited 10 Pre-Mortem tools — then built one that actually works

The 5-minute exercise that saved me 6 months

What's a Pre-Mortem?

The Audit: I checked 10 repos. Best was 2/4.

Results

What I built: premortem-skill

Quick Mode (!pm) — 30 seconds

Full Mode (!pm full) — 2 minutes

How the 4 features work together

1. Base Rates 🎲

2. Bias Circuit-Breaker 🛑

3. L/I Scoring 📊

4. Commitment 🎯

Real example

How to use it

Why this matters

What I built: `premortem-skill`

Quick Mode (`!pm`) — 30 seconds

Full Mode (`!pm full`) — 2 minutes