The Problem

Agentic AI research in healthcare is published across venues and fragmented by modality (radiology, genomics, EHR), application (drug discovery, patient dialogue), and year. A clinician or researcher trying to track the field has no single index. This repo is that index: a curated, categorized list of papers, plus a survey paper and two data-visualization figures summarizing the literature landscape.

What This Does

The entire product is README.md. It is a static, hand-maintained catalog of papers organized two ways: by year (2026, 2025, 2024, 2023) and by category (doctor-facing agents, multi-modal clinical agents, radiology, pathology, cardiovascular imaging, genomics, and more). Each entry links out to the underlying paper.

Two images support the list: landscape.png (a conceptual framework of the agent pipeline, from data perception to application ecosystem) and statistics.png (quantitative analysis of recent literature: top data modalities, dominant technologies, and application domains). The README links to a companion survey paper on TechRxiv and includes a BibTeX citation entry.

There is no code. No functions, no modules, no logic to trace.

How It Is Wired

There is no execution path. This is a documentation repository — three files, one of which is Markdown. Control flow does not apply. The "entry point" is opening README.md in a browser or editor; navigation is via the table of contents anchor links. The only external touchpoints are the hyperlinks to papers and the survey URL. Anyone wanting to extend the list edits README.md directly and opens a pull request, per the README's own instructions.

How To Use It

There is nothing to install or build. No dependency files, no config, no entry point.

git clone https://github.com/moses-y/Awesome-AI-Agents-for-Healthcare

Open README.md and browse the table of contents. To contribute, edit the file and open a PR; the README explicitly welcomes PRs and issue reports for missing papers. No environment variables, no keys, no runtime.

Real-World Use

A research group tracking agentic AI in healthcare uses this as a starting bibliography. A PI assigns a new grad student the radiology section; the student works through the 1.2 Radiology Agents category, follows each paper link, and builds a literature review. The statistics.png figure gives a quick data-driven orientation: textual data dominate, frameworks and multi-agent collaboration are the hot technologies, drug discovery and genomics are emerging frontiers. The survey paper citation is the group's go-to reference for grant applications.

Code Health & Issues

The measured static analysis reports three findings, all from the SDLC axis:

  • Medium – No test files detected. Expected: this is a Markdown-only repo with zero executable code, so tests are not applicable. The finding is technically true but carries no practical weight here.
  • Medium – No CI/CD pipeline detected (no .github/ or CI config). Same caveat: there is no build or test gate to run.
  • Medium – No LICENSE file. This one matters. The repo is a fork of a 1,232-star parent, but without a license, reuse and redistribution rights are legally unclear. Anyone wanting to repurpose the curated list should clarify licensing with the maintainers first.

No secrets were found. The substantive issue is the missing license; the test and CI findings are noise for a documentation repo.

The Bottom Line

This is a useful, well-organized literature index, not a software project — the value is the curation and the linked survey, not code quality. It is a faithful fork of a popular parent repo, and the missing license is the one real blocker for downstream use. Use it as a reference catalog; do not expect tooling, automation, or a codebase to evaluate.