Selects medical imaging architectures and vets model artifacts for license, provenance, and benchmark contamination.
- GitHub stars
- 378
- Files
- 18
- Last commit
- Nov 10, 2026
- License
- MIT
- Allowed tools
- —
What does the model-selection skill do?
model-selection selects medical imaging architectures and vets model artifacts for license, provenance, and benchmark contamination.
Summary generated from the skill's SKILL.md.
About the model-selection skill
Researchers and developers planning medical imaging studies use this skill to select appropriate model architectures and verify the integrity of specific model artifacts. The process begins by framing a task through constraints like modality, dimensionality, and label availability to identify a suitable architecture family, such as U-Net for segmentation or YOLO for detection. It guides the user through a decision tree based on curated reference cards for classification, synthesis, foundation models, and graph neural networks.
Once a specific repository or checkpoint is identified, the skill generates a model dossier to document license details, version pins, and weight provenance. It includes a specialized provenance gate that runs a local script to audit the dossier for major risks, such as benchmark contamination where evaluation data overlaps with training sets. This ensures that researchers do not mistakenly report training scores as independent validation. The skill produces decision notes and JSON audit files to support the documentation of methods and limitations in scientific papers.
What an agent can do with it
- 1Select a segmentation architectureIdentify whether a 3-D U-Net or a foundation model like MedSAM is more appropriate for a specific volumetric segmentation task.
- 2Audit model provenanceCheck if a specific GitHub checkpoint has a documented license and if its training data conflicts with intended evaluation sets.
Written from the skill's SKILL.md. Every use case is checked against a quote from it.
How to install the model-selection skill
Each command copies the skill folder into the client's skills directory. Restart the client or start a new session to load it.Install the model-selection skill in Plugin
/plugin marketplace add Aperivue/medsci-skills /plugin install medsci-modeling@medsci-skills
Install the model-selection skill in Claude Code
~/.claude/skills/model-selection/SKILL.md
$ npx degit Aperivue/medsci-skills/skills/model-selection ~/.claude/skills/model-selectionInstall the model-selection skill in Claude Code (project)
.claude/skills/model-selection/SKILL.md
$ npx degit Aperivue/medsci-skills/skills/model-selection .claude/skills/model-selectionInstall the model-selection skill in Codex
~/.agents/skills/model-selection/SKILL.md
$ npx degit Aperivue/medsci-skills/skills/model-selection ~/.agents/skills/model-selectionInstall the model-selection skill in Gemini CLI
~/.gemini/skills/model-selection/SKILL.md
$ npx degit Aperivue/medsci-skills/skills/model-selection ~/.gemini/skills/model-selectionInstall the model-selection skill in Claude.ai
npx degit Aperivue/medsci-skills/skills/model-selection model-selection && zip -r model-selection.zip model-selection # claude.ai → Customize → Skills → + → Create skill → Upload a skill → model-selection.zip
Automated checks
Ran Nov 15, 2026Full report →Passed our automated checks. That is not a security audit: read the scripts before you let an agent run them. How checks work
- Source repository linked and reachableaperivue/medsci-skills/skills/model-selection is public and reachablegithub.com/aperivue/medsci-skills/tree/HEAD/skills/model-sel ↗Pass
- Open source license (OSI)MITspdx.org/licenses/MIT.html ↗Pass
- Commit or release in the last yearLast commit 2026-11-10github.com/aperivue/medsci-skills/commits ↗Pass
- No download-and-run or credential theft in scripts2 scripts read; no download-and-run, remote eval, base64 execution or credential uploadPass
- No hidden instructions in SKILL.mdNo override phrases, safeguard switches, credential requests or hidden charactersgithub.com/aperivue/medsci-skills/blob/HEAD/skills/model-sel ↗Pass
- Pre-approved tools are scopedNo pre-approved tools; the agent asks before using onegithub.com/aperivue/medsci-skills/blob/HEAD/skills/model-sel ↗Pass
- SKILL.md follows the Agent Skills formatValid name and description; says when to use it; 143 linesgithub.com/aperivue/medsci-skills/blob/HEAD/skills/model-sel ↗Pass
SKILL.md
View on GitHubModel-Selection Skill
Two questions, in this order. Phases 1–3 answer a literature question with a stable answer: which architecture family suits the task. Phases 4–6 answer a provenance question: which concrete artifact — repository, revision, checkpoint — will be run, and what its numbers may claim. The skill writes decision notes and runs one stdlib gate; it never downloads, runs, trains or benchmarks a model, and it describes archetypes, not a live SOTA leaderboard.
Elsewhere: building the repo → /model-scaffold; profiling data and planning preprocessing →
/imaging-data; validation design, metrics, uncertainty, explainability → /model-assessment;
documenting a model you built → /model-card; an LLM/MLLM, including benchmark contamination in
that setting → /mllm-eval; AI-vs-expert reader study → /design-ai-benchmarking.
Workflow
Phase 1 — Frame the question
State the task (classification / segmentation / detection / synthesis / transfer), the modality and dimensionality (2-D vs 3-D volume), the labelled-data scale (events / structures, not just images), label availability (lots / few / unlabelled pool), and the constraints (class imbalance, small structures, interpretability, deployment compute).
Phase 2 — Walk the decision tree, read the family card
Follow ${CLAUDE_SKILL_DIR}/references/index.md (task → constraints → default pick). It routes to
one card; each gives the source paper, core idea, when-to-use, medical-imaging use, reference
implementation, and the typical validation setup for that class.
| Card | Families |
|---|---|
${CLAUDE_SKILL_DIR}/references/classification.md |
ResNet / DenseNet / EfficientNet / Inception / ConvNeXt / ViT / Swin / DeiT |
${CLAUDE_SKILL_DIR}/references/segmentation.md |
U-Net / 3-D U-Net / V-Net / Attention & Residual U-Net / nnU-Net (+ ResEnc, MedNeXt, STU-Net) / SegResNet / Swin-UNETR / Mask R-CNN |
${CLAUDE_SKILL_DIR}/references/detection.md |
nnDetection / Faster R-CNN + FPN / Mask R-CNN / RetinaNet / YOLO / DETR |
${CLAUDE_SKILL_DIR}/references/synthesis.md |
Pix2Pix / CycleGAN / SPADE / diffusion (DDPM, latent) / VAE / fastMRI reconstruction |
${CLAUDE_SKILL_DIR}/references/foundation_models.md |
SAM / MedSAM / MedSAM2 / nnInteractive / VISTA3D / TotalSegmentator / SegVol / BiomedCLIP / DINO / MAE / SimCLR / MoCo |
${CLAUDE_SKILL_DIR}/references/graph.md |
GCN / GraphSAGE / GAT / GIN / BrainGNN for connectomes and population graphs (PyTorch Geometric / DGL; not scaffolded by /model-scaffold) |
Never recommend an architecture for a modality or data scale it does not suit (a from-scratch ViT on a few hundred images, 2-D slices for a volumetric structure) — the decision-tree constraints exist to prevent exactly that. If asked for "the best" model, say the zoo is a curated archetype map, not a current SOTA ranking.
Phase 3 — Write the architecture decision note
Write decisions/architecture_choice.md: the task, the chosen architecture, its source paper
(mandatory — the Methods cite it), the reason against the constraints, the runner-up and why not,
and the matching /model-scaffold template. Never invent a benchmark number or paper claim: cite it
(verify via /search-lit), or write [VERIFY] and ask. A reference without a /search-lit-confirmed
DOI/PMID is marked [UNVERIFIED - NEEDS MANUAL CHECK].
Phase 4 — Write the model dossier for the concrete artifact
When a concrete candidate exists (a GitHub repo, a Hugging Face checkpoint, a paper's released
weights), write model_dossier.json recording what is known; leave unknowns unstated, never
guessed — the gate turns an absence into a finding.
{
"model": "OrganSeg-3D v2.5.1",
"source": {"kind": "github", "url": "...", "version": "v2.5.1", "commit": "abc1234"},
"licence": {"spdx": "Apache-2.0", "verified_from": "LICENSE at commit abc1234"},
"intended_use": "research",
"weights": {"pretrained": false},
"task": {"model": "3d_ct_organ_segmentation", "study": "3d_ct_organ_segmentation"},
"reported_validation": [{"dataset": "ExampleBench", "metric": "Dice", "source": "J Ex 2021"}],
"developed_on": ["ExampleBench"],
"evaluation_arms": [{"name": "external", "dataset": "OtherCohort-2026"}],
"hardware": {"claimed": "any CUDA GPU", "verified_on": "GTX 1080 Ti", "verified": true}
}
Read each field from the artifact, not from memory:
- Licence — only from the licence file at the pinned revision, and record which file
(
verified_from). Never from a README badge, a model-card summary, or memory; an unstated licence isLICENCE_UNSTATED, never "probably MIT". developed_on— the paper's own account of where the method was built and tuned. It is the field people skip and the one the gate needs: a method that won a challenge was tuned on it.hardware.verified— true only after it actually ran. A support matrix is a different claim: a CUDA capability one compiler accepts may still be refused by another in the same stack.
Phase 5 — Gate the dossier
python3 ${CLAUDE_SKILL_DIR}/scripts/check_model_provenance.py --dossier model_dossier.json \
--out qc/model_provenance.json --strict
Stdlib-only and network-free: nothing is fetched and no licence resolved online.
| Verdict | Severity | Fires when |
|---|---|---|
BENCHMARK_PROVENANCE_CONFLICT |
Major | an evaluation arm uses a dataset the model was developed or tuned on |
EVAL_DATA_IN_TRAINING |
Major | an evaluation arm's dataset is inside the pretraining corpus |
LICENCE_UNSTATED |
Major | no licence recorded — not the same as a permissive one |
LICENCE_INCOMPATIBLE |
Major | non-commercial / research-only licence under commercial or deployment intent |
WEIGHTS_PROVENANCE_UNKNOWN |
Major | pretrained weights whose training corpus is not stated |
DEVELOPED_ON_UNSTATED |
Major | the developed_on key is absent, so the conflict check could not run |
EVALUATION_ARMS_UNSTATED |
Major | the evaluation_arms key is absent, so neither provenance check could run |
INTENDED_USE_UNSTATED |
Major | intended_use is absent, so licence compatibility could not be checked |
TASK_MISMATCH |
Minor | the model's task is not the study's task |
NO_VERSION_PIN |
Minor | no commit, tag or revision |
VALIDATION_UNREPORTED |
Minor | no reported validation (dataset + metric + source) |
HARDWARE_UNVERIFIED |
Minor | hardware support claimed but never executed |
LICENCE_UNVERIFIED |
Minor | a licence is named but not the file it was read from |
The gate flags a relationship, not a reputation: developed_on: ExampleBench passes cleanly as
long as no evaluation arm uses ExampleBench (the clean fixture proves this). Dataset names match as
token sequences with a small family-alias table — MSD Task09 Spleen matches MSD,
MS Cohort 2026 does not — never by substring.
An explicit empty list ("developed_on": []) states "none" and is not a finding; an absent key is.
intended_use must be research, commercial or clinical_deployment, and developed_on /
weights.trained_on must be lists of strings; anything else is an input error (exit 2), never a pass.
Known limits. LICENCE_INCOMPATIBLE recognises non-commercial / research-only markers in the
licence string (NC, non-commercial, research-only, in SPDX or spaced form). A free-text licence
that restricts use without those markers (e.g. "academic use only"), or a copyleft licence under
closed deployment, is not detected: read the licence file and record the conclusion yourself.
Phase 6 — Turn a Major into a study decision
A BENCHMARK_PROVENANCE_CONFLICT rarely means abandoning the model (it is often the best-engineered
option precisely because it was tuned hard). It changes what the arm may claim:
- Report that arm as a demonstration that the pipeline runs end to end, not as evidence the method works.
- Put the evidential weight on an arm whose data post-dates the model, and say so with dates.
- State the conflict in Methods and Limitations rather than leaving a reviewer to find it.
Never report a flagged arm as independent validation. An EVAL_DATA_IN_TRAINING arm is different in
kind: it produces a training-set score and cannot be reported as validation at all. Write this
arm-by-arm decision into the study record.
Outputs and hand-off
decisions/architecture_choice.md— the architecture decision note (Phase 3).model_dossier.json— the provenance record (Phase 4);qc/model_provenance.json— the audit (Phase 5).- The arm-by-arm decision (Phase 6).
Carry them to /model-scaffold (instantiate the template), /model-assessment (arm and validation
design, what each arm may claim, metrics), /model-card (provenance section), and /write-paper
(Methods cite the source paper; Limitations state any conflict). Gate regression:
bash ${CLAUDE_SKILL_DIR}/scripts/check_model_provenance_challenge/verify.sh and
bash ${CLAUDE_SKILL_DIR}/tests/test_model_provenance.sh.
Files in this skill
18model-selection skill questions
How do I install the model-selection skill in Claude Code?
Run npx degit Aperivue/medsci-skills/skills/model-selection ~/.claude/skills/model-selection. Claude Code loads skills from ~/.claude/skills (all projects) or .claude/skills (one project) and uses a skill when its description matches the task.
Which clients can use the model-selection skill?
Any client that reads the Agent Skills format: Claude Code, Codex, Gemini CLI and claude.ai (upload as a ZIP). Scripts in the skill run only where the client allows code execution.
What kind of architectures can this skill help select?
The skill covers a wide range of families including CNNs (ResNet, ConvNeXt), ViTs (Swin, DeiT), segmentation models (U-Net, nnU-Net), detection models (YOLO, DETR), foundation models (SAM, MedSAM), and Graph Neural Networks (GCN, GAT).
Does this skill download or run models?
No, the skill describes archetypes and writes decision notes and dossiers; it never downloads, runs, trains, or benchmarks a model.
How does the skill handle model license verification?
It requires reading the license directly from the pinned revision's license file and recording the source; it does not rely on README badges or memory.