- Home
- Skills
- Agents & Orchestration
- medical research council
Works with the AI Agents you already use
medical research council
A multi-agent review board that pressure-tests medical research protocols and grants through anonymous peer review.
Free
medical research council
Example session with this skill installed
Review my protocol for a randomized controlled trial comparing a new AI-driven insulin pump vs standard care in Type 1 Diabetes patients. --type protocol
- Read your context and instructions
- Compiled the medical research council
I have completed the Medical Research Council review for the insulin pump RCT protocol. The session folder has been created at ./council-sessions/2024-05-20-ai-insulin-pump-rct/.
The council has issued a verdict of
Conditional Proceed (Score: 7.2/10). A Safety & Ethics warning was triggered regarding the emergency override latency in the intervention group.
Verdict Summary
| Option | Weighted Score | Rank | Veto Status |
|---|---|---|---|
| A: Full RCT as drafted | 7.2 | 1 | Warning |
| B: Pilot/Feasibility study | 8.1 | 2 | Clear |
| C: Do not proceed | 4.5 | 3 | Clear |
Key Arguments
Methodologist: The sample size calculation (n=120) is underpowered if the primary endpoint is HbA1c reduction at 3 months; recommend extending to 6 months.
Ethics & Regulatory: High-risk of "automation bias" among participants. The protocol lacks a defined manual override protocol for hypoglycemic events.
Clinical Lead: Strong clinical utility. The current comparator (standard care) is appropriate, but the recruitment timeline is overly optimistic for a single-center study.
Strongest Objection
The
Evidence & Safety seat flagged that the current pump firmware has not yet received its Class IIb SaMD equivalence, which may delay IRB approval by at least 6 months.
Next steps
- Review the specific safety objections in
council-sessions/2024-05-20-ai-insulin-pump-rct/round-1.md. - Update the "Emergency Procedures" section of your protocol to address the manual override latency.
- Run
medical-research-council rerun ./council-sessions/2024-05-20-ai-insulin-pump-rct/once the revisions are saved.
Connects securely to your tools. The creator never sees your data.
About this skill
Medical research is high-stakes, yet most protocols and grant ideas are reviewed too late in the cycle. This skill provides a specialist review board that pressure-tests research topics, study designs, and MedTech projects before you commit resources or patients.
What it does
- Multi-agent peer review Spawns five independent specialist sub-agents (Methodologist, Ethicist, Editor, etc.) to analyze your project.
- Anonymous critique Facilitates a second round where seats review each other's blinded analyses to catch bias or blind spots.
- Evidence verification Cross-references all claims against PubMed and ClinicalTrials.gov to prevent hallucinated citations.
- Objective scoring Calculates weighted 1-10 scores and applies a hard Safety & Ethics veto to determine project fundability and feasibility.
- Automated reporting Generates a comprehensive
verdict.mdand an HTML report with ranked options and specific objections.
How it works
- The Brief Provide your research question, protocol, or grant text for the agent to classify and frame as a PICO(T) decision.
- Evidence Pack The agent searches literature and registries to build an
evidence.mdfile of existing systematic reviews and ongoing trials. - The Council The agent runs independent Round 1 analyses followed by a Round 2 anonymous peer review.
- Scoring & Verdict A dedicated Python script processes weights and vetoes to produce the final ranking and verdict report.
Frameworks & tools
This skill utilizes Python 3 scripts for scoring and anonymization. It integrates with PubMed, Consensus, and ClinicalTrials.gov MCP tools for literature retrieval.
Why this beats prompting it yourself
Standard prompting often leads to sycophancy or "echo chamber" feedback. This skill enforces strict independence between agents and uses a secondary blinding round to surface genuine technical disagreements and regulatory red flags.
Use cases
- Grant preparation Identify likely reviewer objections and feasibility gaps before submission.
- Protocol design Optimize sample size, endpoints, and bias mitigation strategies for clinical trials.
- MedTech validation Determine the appropriate regulatory pathway and SaMD classification for healthcare AI.
- Manuscript review Conduct a pre-submission "roast" to select the best target journal and fix methodological flaws.
Known limitations
Does not diagnose or treat patients. Not for hospital administration decisions. Requires an environment capable of running Python 3 scripts for full reporting features.
Find more skills like this on the Claude Skills Marketplace, the Codex Skills Marketplace and the Gemini Skills Marketplace, or browse Claude Skills for Agents & Orchestration.
Does not diagnose or treat patients. Not for hospital administration decisions. Requires Python 3.8+ to run scoring and anonymization scripts.
How to use
- 1
Connect your agent to the Agensi MCP
Copy the prompt and paste it into Claude, ChatGPT, Cursor or any agent that supports MCP. It sets up the connection for you. Manual setup
- 2
Sign in to Agensi
Your agent shows a sign-in link. Sign in and click Allow.
- 3
Ask for what you need
Your agent loads this skill and uses it on your task. It always asks before buying a skill.
Reviews
No reviews yet
Be one of the first to try it. Every listed skill goes through the trust checks below.
Security scanned
Checked by our 8-point scan before listing
Fresh listing
Recently published to Agensi
Free to use
No account required to browse
Trust & safety
Security scan
Security scan passed 6 days ago
- Free with an Agensi account