// SCAN REPORT

MedSci Agent

28 tools. 6 can modify or destroy data without limits.

1 destructive tool with no built-in limits. Policy required.

Last updated: 11 Jun 2026

6 can modify or destroy data

22 read-only

28 tools total

Community server · catalogue entry verified 11/06/2026

How to control MedSci Agent ↓

TOOL MAP

What MedSci Agent exposes to your agents

Read (22) Write / Execute (5) Destructive / Financial (1)

Running more than one server? Check your whole stack →

MOST DANGEROUS TOOLS

Critical Risk

The most dangerous MedSci Agent tools

Destructive · High sandbox_teardown While stopping a sandbox is reversible (Execute category), the tool's ability to remove/delete Docker containers crosses into Destructive territory because deletion cannot be undone. Execute · Medium differential_expression This is an Execute tool because it runs complex statistical and computational operations (differential expression analysis algorithms) whose outputs depend entirely on input parameters and data. Execute · High sandbox_run_job This tool executes arbitrary commands in a Docker container. Execute · High sandbox_prepare This tool provisions and configures a Docker container environment, which constitutes executing external infrastructure operations.

6 of MedSci Agent's 28 tools can modify, destroy, or commit something on every call — and an agent calls them with no built-in limits.

POLICY

How to control MedSci Agent

PolicyLayer is an MCP gateway — it sits between your AI agents and MedSci Agent, and nothing reaches the server without passing your rules. These are the rules we recommend:

Deny destructive operations

{
  "sandbox_teardown": {
    "deny_if": [
      {
        "conditions": [],
        "on_deny": "Blocked by default. Requires approval."
      }
    ]
  }
}

Destructive tools should never be available to autonomous agents without human approval.

Cap read operations

{
  "parse_fasta": {
    "limits": [
      {
        "counter": "parse_fasta_per_minute",
        "window": "minute",
        "max": 60,
        "scope": "grant"
      }
    ]
  }
}

Controls API costs and prevents retry loops from exhausting upstream rate limits.

Create a free account and register MedSci Agent — nothing to install.
Add these rules — paste them, or build them visually. Tune the limits to your setup.
Point your MCP client (Claude, Cursor, anything) at your gateway URL.

ENFORCE POLICY ON MEDSCI AGENT →

Free to start. No card required.

FULL CATALOGUE

All 28 MedSci Agent tools

DESTRUCTIVE 1 tools

Destructive sandbox_teardown Stop or remove a Docker sandbox. By default only stops; set remove to true to also delete.

EXECUTE 5 tools

Execute differential_expression Run differential expression analysis between cell groups. Identifies genes that are significantly up- or down- Execute sandbox_run_job Execute a command in an existing Docker sandbox and capture stdout, stderr, exit code, duration, and artifact Execute cluster_cells Cluster cells in a preprocessed single-cell dataset using Leiden or Louvain community detection. Returns clust Execute preprocess_omics Preprocess a single-cell RNA-seq dataset: filter cells/genes, normalize, log-transform, and identify highly va Execute sandbox_prepare Create or verify an isolated Docker sandbox for the workspace. Optionally apply a container template and netwo

READ 22 tools

Read parse_fasta Parse a FASTA file and return sequence metadata: IDs, names, descriptions, lengths, and sequence previews. Read acquire_documents Acquire policy-compliant scientific document text from DOI/PMID/PMCID/URL targets with provenance and content- Read analyze_medical_image Analyze a medical image (X-ray, CT, pathology slide, dermatology photo) using MedGemma Read analyze_molecule Analyze a molecule from its SMILES string. Returns physicochemical properties: molecular weight, LogP, H-bond Read analyze_sequence Analyze a protein or DNA sequence: compute length, amino acid composition, molecular weight. For DNA, can also Read fetch_abstract Fetch abstract text and key metadata for a PubMed article by PMID. Returns abstract text, title, MeSH terms, j Read lipinski_filter Check if a molecule passes Lipinski Read molecular_similarity Compute Tanimoto similarity between two molecules using Morgan fingerprints (ECFP4). Score ranges 0-1, where > Read predict_admet Predict ADMET (Absorption, Distribution, Metabolism, Excretion, Toxicity) properties for a molecule. Returns A Read predict_structure Retrieve an AlphaFold-predicted protein structure by UniProt accession. Returns confidence scores (pLDDT), pre Read predict_toxicity Predict compound toxicity Read read_h5ad Read an H5AD (AnnData) file and return metadata: number of observations, variables, column names. Use this as Read sandbox_fetch_artifact Read a generated artifact or log file from the host-visible workspace path. Enforces path safety checks and si Read sandbox_status Return the high-level state of a Docker sandbox (exists, running, or stopped). Read search_and_analyze Performs deep literature analysis using PaperQA2. Accepts either DOI/PMID paper identifiers (with internal NCB Read search_chembl Search the ChEMBL database for bioactive molecules. Find compounds by target, molecule name, or similar struct Read search_clinical_trials Search ClinicalTrials.gov study metadata by condition, intervention, or NCT identifier. Read search_openalex Search OpenAlex for scholarly metadata across disciplines. Returns titles, authors, citation counts, open-acce Read search_pubmed Search PubMed for biomedical literature metadata. Returns titles, authors, journal, publication date, article Read search_uniprot Search the UniProt protein database. Find proteins by name, gene, organism, or accession. Returns protein func Read gene_set_enrichment Perform gene set enrichment analysis on a list of genes. Queries Enrichr or GSEA against MSigDB, GO, KEGG path Read resolve_identifier_to_sources Resolve a DOI/PMID/PMCID to candidate source URLs with provenance and confidence metadata.

EXPLORE

Related servers

Other MCP servers with similar tools — same risk classification, starter policies for each.

FAQ

Questions about MedSci Agent

Can an AI agent delete data through the MedSci Agent MCP server? +

Yes. The MedSci Agent server exposes 1 destructive tools including sandbox_teardown. These permanently remove resources with no undo. PolicyLayer blocks destructive tools by default so they never reach the upstream server.

How many tools does the MedSci Agent MCP server expose? +

28 tools across 4 categories: Destructive, Execute, Read, Write. 22 are read-only. 6 can modify, create, or delete data.

How do I enforce a policy on MedSci Agent? +

Register the MedSci Agent MCP server in PolicyLayer, apply the suggested rules above (adjust the limits to your use case), and point your AI client at the PolicyLayer proxy URL instead of the server directly. Your agents keep the same tools; PolicyLayer evaluates every call against policy before it executes. Nothing to install, live in minutes.

Enforce policy on every MedSci Agent tool call.

Deterministic rules across all 28 MedSci Agent tools. Per-identity grants. Full audit log. Live in minutes. Nothing to install.

ENFORCE POLICY ON MEDSCI AGENT →

Free to start. No card required.

28 MedSci Agent tools catalogued and risk-classified — across an index of 43,000+ MCP servers.