Our Mission
AI Toolkit
Overview DAG Studio SPIFD2 Workspace TARGET Checklist ROBINS-I RWE MCP Registry AI Fluency
Reading List
This Week's Picks Archive How It Works
Blog
About
Contact
Open source · Apache-2.0 Official MCP Registry pip install robins-i-mcp

Risk of Bias in Non-Randomized Studies

ROBINS-I MCP

An MCP server implementing ROBINS-I V2 for follow-up cohort studies. The agent reads and cites. The algorithm decides.

// illustrative example: a completed assessment Illustrative example of a ROBINS-I V2 risk-of-bias assessment rendered by ROBINS-I MCP, using synthetic data: the assessed result and outcome, the target trial specification, an overall judgement of Moderate, and a strip showing each of the six domain judgements. Illustrative example · synthetic data, not a real publication

In Brief

Answers from the agent. Judgements from the algorithm.

Each ROBINS-I domain has a published decision flowchart: the signalling-question answers pick the path, and the path gives the judgement. Here the agent only answers, and every answer carries a quote. Walking the flowchart is code, so two assessors with the same answers reach the same verdict.

// how an assessment runs

LLM agent

Answers questions, cites text

Reads the paper, answers the signalling questions the algorithm actually reaches, and supplies the evidence for each one.

ROBINS-I MCP server · 9 tools
Ingestparse_document · parse_pmcid
Setupset_prespecified_confounders · specify_result
Assessget_spec · assess_result · submit_answers
Renderrender_report
Reviewexport_robvis

Deterministic. Holds the seven domain algorithms as explicit edge graphs and walks them.

The unit of assessment is one numerical result, not one paper: a study reporting three outcomes across two analyses is six assessments.

01 Ingest the paper and its supplement via MCP

parse_document builds a source-tagged map and returns a survey of fifteen evidence cues: where each domain’s evidence tends to live, and which absences already look real. The supplement matters: the target-trial specification often lives only there.

02 Two gates, before anything is scored via MCP

set_prespecified_confounders supplies the review’s list of important confounders. Domain 1 asks whether all of them were controlled, and “important” is the reviewer’s judgement, not the paper’s covariate table. So the server refuses to score domain 1 without it rather than silently substituting one for the other. Then specify_result settles C4, which selects which of domain 1’s two question sets even exists.

03 Answer one domain at a time agent

assess_result returns a scaffold per domain, never one flat rubric; most of the forty signalling questions are unreachable on any given path. On the reference assessment, twenty-four were never reached, and the algorithm’s branching decided that, not the assessor.

04 Evidence is enforced, not requested server-side

Every quote is resolved to a character offset in the ingested text; an unfound quote is rejected, with the nearest actual text returned so it can be corrected. A claim that the paper is silent must name a search cue, and the server runs the search: prose asserting that you looked is refused, because it cannot be audited.

05 The algorithm decides deterministic

No tool accepts a domain judgement as an input. The six domain algorithms and the overall algorithm are explicit edge graphs, traced arrow by arrow from the published flowcharts, and the verdict is the traversal. A reviewer may override it, with a recorded justification and both values shown.

06 A stamped assessment, and a ratification queue artifact of record

The report carries a provenance stamp bound to the ingested text and to a fingerprint of the traced algorithms, so a later correction is detectable in work produced under the earlier version. Anything resting on reviewer judgement rather than the paper stays in a ratification queue: the assessment is explicitly not final until a human signs it off.

Open source · Apache-2.0 pip install robins-i-mcp com.blackswancausallabs/robins-i-mcp in the MCP Registry Independent implementation, not endorsed by the ROBINS-I development group. The 20 November 2025 release of the tool is a draft, and the algorithms here were traced by hand from its figures and remain externally unverified; every report says so on its face.