ClawHub

Medical Research Toolkit

v1.0.0

Query 14+ biomedical databases for drug repurposing, target discovery, clinical trials, and literature research. Access ChEMBL, PubMed, ClinicalTrials.gov, O...

⭐ 3· 1.9k·22 current·24 all-time
byPascal Bro@pascalwhoop
MIT-0
Download zip
LicenseMIT-0 · Free to use, modify, and redistribute. No attribution required.
Security Scan
VirusTotalVirusTotal
Benign
View report →
OpenClawOpenClaw
Benign
high confidence
✓
Purpose & Capability
Name/description match the content: SKILL.md and reference files consistently document queries against many biomedical databases via a single unified MCP endpoint. No unrelated credentials, binaries, or config paths are requested.
ℹ
Instruction Scope
All runtime instructions are curl/json-rpc examples that POST queries to https://mcp.cloud.curiloo.com/tools/unified/mcp (or optionally to a local 'medical-mcps' service). The instructions do not ask the agent to read local files or hidden credentials, but they do direct potentially sensitive query text to an external third-party endpoint — this is expected for the stated purpose but has privacy implications (do not send patient-identifiable data).
✓
Install Mechanism
No install spec or code is included in the skill bundle (instruction-only). The SKILL.md mentions 'pip install medical-mcps' as a local option, but that is an external package installation separate from the skill; the skill itself writes nothing to disk.
✓
Credentials
The skill declares no required environment variables or credentials. Where APIs require keys (OMIM, optional OpenFDA/NCI), the docs show passing them as API parameters; the requested access is proportionate to the documented APIs. No unrelated secrets are requested.
✓
Persistence & Privilege
The skill does not request always:true or elevated persistent privileges, and it does not modify agent/system configs. It is user-invocable and may be called autonomously by the agent (normal default).
Assessment
This skill is documentation and curl examples for a third‑party unified biomedical API; it appears internally consistent with that purpose. Before using it, verify the external endpoint (https://mcp.cloud.curiloo.com) and the upstream source (README links to a GitHub repo) so you trust who will receive your queries. Never send patient‑identifiable information or sensitive clinical data to the public endpoint. If you prefer local control, inspect the 'medical-mcps' package source on PyPI/GitHub before installing and consider running the service locally behind your network controls. Note small doc inconsistencies (OpenTargets uses EFO vs MONDO ids in different files) — validate returned IDs when chaining lookups. If you need a higher assurance review, ask for the upstream service code or a signed API/owner identity before sending sensitive queries.

Like a lobster shell, security has layers — review code before you run it.

latestvk976bmdrsc25d3m5dnhffggpxx81a04f

License

MIT-0
Free to use, modify, and redistribute. No attribution required.
Termshttps://spdx.org/licenses/MIT-0.html

Comments