write-report-adapter

name	write-report-adapter
description	Write a chest X-ray report grounded in the trained MedImageInsight (MI2) adapter. Invoke whenever drafting a CXR report when the adapter is available — get vision-grounded per-finding positive/possible/negative calls, then delegate the prose to `write-report-simple`.
argument-hint	- image_path: required path to the CXR image (PNG or JPEG) - head: optional adapter head — "mlp" (default) or "svm"

write-report-adapter

Two-step skill: (1) call the trained MI2 adapter to get per-finding positive/possible/negative calls for the image, then (2) hand those off to the existing write-report-simple skill to produce the actual report. The report itself never mentions MI2, the adapter, or any classifier — it reads as the radiologist's direct read of the image.

Budget

• Tool calls: 1 — exactly one mcp_medimageinsig_adapter_classify call per image.
• Output: ≤ 200 words for the full report.
• Do not re-invoke skills already used in this turn.
• Do not retry tool calls on success.

How to use

1. Resolve the image to an absolute path. The MCP tool runs in a different working directory than the agent. If image_path is not absolute, resolve it (e.g. via realpath) before calling the tool. If you can't resolve it, ask the user.

2. Get calls from the adapter.

   • Call mcp_medimageinsig_adapter_classify with file_path = the absolute path from step 1, and head = the caller's head if provided (otherwise omit; defaults to mlp).
   • The adapter returns {label: "positive" | "possible" | "negative"} — already bucketed using per-class operating points fit during training (Youden-J for rule-in, sens90 for rule-out). No further thresholding.

3. Hand off to write-report-simple. Invoke the write-report-simple skill (already in the workspace skill drawer at .github/skills/write-report-simple/). Pass:

   • findings: the list of labels with call "positive" or "possible", using the original adapter label names. Omit negatives.

   The downstream skill derives confidence from the positive/possible mix. The adapter's possible label maps to hedged language (possible, differential includes, cannot exclude) in the report.

4. Return what write-report-simple returns. Do not re-wrap, re-format, or annotate the report. The downstream skill emits the <confidence> and <report> blocks; pass them through unchanged.

Critical rule

The report must not mention MI2, the adapter, classifier scores, or any model output. That's the responsibility of write-report-simple (which has the rule baked in), but it's worth restating here: this skill exists to ground the report in vision-model evidence, not to surface that evidence in the prose. If you want to attribute or caveat the model in the chat reply outside the <report> tags, that's fine.