Academic AIO
Audits medical-AI papers for AI-search discoverability, with a pass/fail checker
Test report
- Verdict
- Tested · Works
- Score
- Tested
- Jul 16, 2026
- Environment
- Claude Code 2.x (agent harness)
- Upstream re-checked
- Aug 10, 2026 · 82c9c02
Task: prepare a title, Key Points box, and abstract for submission to Radiology — a transformer detecting skull fractures on non-contrast head CT. Baseline (without skill) produced a reasonable-looking draft, but the skill's built-in detector rejected it: `check_summary_box.py --journal radiology --strict` returned EXIT=1, verdict NONCONFORMANT — 'found 4 bullets, expected 3' (RSNA requires exactly 3) plus a soft-hit one_claim_per_bullet on a bullet with a semicolon. The skill's version passed: EXIT=0, verdict CONFORMANT. Text measurements: baseline had 0 mentions of '95% CI', 0 numerical outcomes (AUC/sensitivity/specificity), 0 reporting guideline anchors, 0 code-availability with DOI; the skill's version had — 5 / 5 / 2 / 1 respectively. All 12 files referenced by the body (scripts/check_summary_box.py, references/summary_box_specs.json, references/checklists/AIO_GENERAL.md, templates/aio_audit_checklist.md.j2, etc.) actually exist, and three repository bundle tests were run and passed (test_summary_box: 7 passed 0 failed; test_validate_schema and test_batch_metadata_audit: ALL PASS). Scripts are stdlib-only, no network or environment reading — grep for urllib|requests|socket|subprocess|os.environ|token|secret only matched 'https://schema.org' and regex ORCID.
Scored on four weighted criteria — install, triggering, output vs. baseline, docs. How scoring works
- Installs cleanly 5/5
- Triggers reliably 5/5
- Output vs. baseline 9/10
- Docs & honesty 4/5
What Academic AIO does
Reviews titles, abstracts, journal summary boxes, READMEs, CITATION.cff and Hugging Face cards for medical-AI papers so Perplexity, Elicit, Consensus and SciSpace cite them accurately, combining GEO principles with TRIPOD+AI/CLAIM/STARD-AI reporting anchors. Triggers when drafting or reviewing a manuscript, preprint, or code release for venues like Radiology, Lancet Digital Health or npj Digital Medicine. Ships Python checkers that validate summary-box format, JSON-LD schema and repo metadata rather than relying on model judgment alone.
How to install Academic AIO
git clone https://github.com/Aperivue/medsci-skills.git
mkdir -p ~/.claude/skills
cp -r medsci-skills/skills/academic-aio ~/.claude/skills/academic-aio
Skills live in ~/.claude/skills/ (global) or .claude/skills/
(per-project). Restart Claude Code after installing.
Commands — how to trigger Academic AIO
-
/academic-aioAudits medical-AI papers for AI-search discoverability, with a pass/fail checker
It also activates on plain-language prompts like these:
-
Check if my Radiology submission's Key Points box has too many bullets -
Help me add TRIPOD+AI reporting anchors to my medical AI manuscript abstract -
Audit my preprint's summary box so Perplexity and Elicit cite it accurately
Frequently asked questions
- Is the Academic AIO skill free?
- Yes. The skill itself is free from Aperivue/medsci-skills. SkillProof publishes the install command and an independent test verdict at no cost.
- Does Academic AIO work with Claude Code?
- We tested it with Claude Code 2.x (agent harness) on Jul 16, 2026. Verdict: Tested · Works. Task: prepare a title, Key Points box, and abstract for submission to Radiology — a transformer detecting skull fractures on non-contrast head CT. Baseline (without skill) produced a reasonable-looking draft, but the skill's built-in detector rejected it: `check_summary_box.py --journal radiology --strict` returned EXIT=1, verdict NONCONFORMANT — 'found 4 bullets, expected 3' (RSNA requires exactly 3) plus a soft-hit one_claim_per_bullet on a bullet with a semicolon. The skill's version passed: EXIT=0, verdict CONFORMANT. Text measurements: baseline had 0 mentions of '95% CI', 0 numerical outcomes (AUC/sensitivity/specificity), 0 reporting guideline anchors, 0 code-availability with DOI; the skill's version had — 5 / 5 / 2 / 1 respectively. All 12 files referenced by the body (scripts/check_summary_box.py, references/summary_box_specs.json, references/checklists/AIO_GENERAL.md, templates/aio_audit_checklist.md.j2, etc.) actually exist, and three repository bundle tests were run and passed (test_summary_box: 7 passed 0 failed; test_validate_schema and test_batch_metadata_audit: ALL PASS). Scripts are stdlib-only, no network or environment reading — grep for urllib|requests|socket|subprocess|os.environ|token|secret only matched 'https://schema.org' and regex ORCID.
- What is the Academic AIO SkillProof Score?
- 9.2/10 — installs cleanly 5/5, triggers reliably 5/5, output vs. baseline 9/10, docs & honesty 4/5.
- How do I install Academic AIO?
- Copy the install command from this page, run it in your terminal, and restart Claude Code. Skills live in ~/.claude/skills/ (global) or .claude/skills/ inside a project.
- Can I use Academic AIO with Cursor, Copilot, Gemini CLI, Codex or other AI tools?
- The SKILL.md format is native to Claude (Claude Code, Desktop, claude.ai). The instructions inside adapt to other assistants: Cursor rules, GitHub Copilot instructions, Windsurf rules, Custom GPTs, AGENTS.md for OpenAI Codex, and GEMINI.md for Google Gemini CLI — our conversion guides cover each, and the free converter on the tools page does the wrapping for you.