Ready to supercharge your AI agents?

Install any skill in seconds. Free to start, no credit card required.

Write-Protocol Skill · aperivue | DecimalAI

scienceby aperivue

Write-Protocol Skill

aperivue/write-protocol

Source aperivue/medsci-skills ↗License MIT

IRB/ethics committee research protocol generator. Produces 4 core sections (Background, Study Design, Sample Size, Statistical Plan) with full prose, plus 6 skeleton sections with TODO markers for institution-specific content. Integrates outputs from design-study, calc-sample-size, and search-lit.

SkillScore32Maintained

Automatic

Clinical / Healthcare pack

Last verified 9/25/2026

What it changes

Passed 6 of 18 test cases with the skill, and 2 of 18 without it.

Without skill14%

With skill32%

Checks it passed that the model alone missed

Section 5 contains a [TODO] placeholder marker.
Section 8 contains a markdown table for milestone schedule.
Section 6 includes [TODO] markers for data retention period.

and 8 more in the full report

5 fixed2 passed either way14 still failing1 broke

Pass-rate lift: +18 pp
Tokens: +37%
Agent turns: 0%
Response time: -7%

Model gemini-3.6-flashMeasured 9/25/2026v2 · latest is v37 of 22 cases passed with skill · 18 comparable

Tokens, turns and time describe resource use, not a dollar price.

This benchmark measured an earlier version of the skill.

A newer attempt on gemini-3.6-flash on 10/3/2026 did not produce a publishable comparison (not comparable). The snapshot above is the previously measured run.

Use this skill

Save it into your project's skills folder. Claude Code picks it up in its next session.

curl -fsSL --create-dirs -o .claude/skills/aperivue-write-protocol/SKILL.md https://app.decimal.ai/s/aperivue-write-protocol/SKILL.md

Want it pinned and updatable? pip install "decimalai>=0.15", then decimalai skills add aperivue-write-protocol, and start Claude Code with claude --plugin-dir "$(git rev-parse --show-toplevel)/.agents"

1 · One-time — install the CLI (docs ↗)

pip install "decimalai>=0.15"

2 · Get the skill — writes .agents/skills/aperivue-write-protocol/ and pins it in .decimal/skills.lock

decimalai skills add aperivue-write-protocol

No CLI? Get the same file — curl · raw file

curl -fsSL --create-dirs -o .claude/skills/aperivue-write-protocol/SKILL.md https://app.decimal.ai/s/aperivue-write-protocol/SKILL.md

Raw file: /s/aperivue-write-protocol/SKILL.md ↗ · machine-readable summary at /s/aperivue-write-protocol.json ↗

Safety checks: 3 of 3 passed

Code scan CleanIntent PassedContent Clean

See it in action

One of 5 test cases the skill fixed. Same model, same prompt, with and without the skill.

1 of 5

Prompt

I need an ethics-ready medical research protocol document generated for my study on non-alcoholic fatty liver disease progression. My primary hypothesis is that baseline serum ferritin levels predict 5-year progression to liver fibrosis in adult patients diagnosed with NAFLD. This is a retrospective cohort study using electronic health records from major urban medical centers between 2015 and 2023. The primary endpoint is biopsy-confirmed or transient elastography-confirmed hepatic fibrosis, while secondary outcomes include overall mortality and metabolic complications. Please assemble a comprehensive protocol markdown file that drafts full scientific prose for background, methodology, power calculations, and analysis plans, while leaving structured placeholder markers for site-specific items like local IRB forms, budgeting, and institutional timelines.

What was checked, without and with the skill
Checked	Without skill	With skill
Section 1 Background and Rationale contains no bullet points.	Missed	Passed
Output contains all 10 numbered protocol sections from Section 1 through Section 10.	Missed	Passed
Section 5 contains a [TODO] placeholder marker.	Missed	Passed

case-01 · gemini-3.6-flash · 9/25/2026 · answers are capped and redacted

What's inside

The instructions your model reads when the skill loads.

SKILL.md116 lines3 files

Write-Protocol Skill

Draft the scientific core of an IRB/ethics protocol and leave institution-specific sections as [TODO] skeletons. Read both reference files before generating a protocol draft:

${CLAUDE_SKILL_DIR}/references/protocol_template.md -- the 10-section structure, per-section guidance, and the [TODO] blocks for Sections 5-9
${CLAUDE_SKILL_DIR}/references/ethics_checklist.md -- jurisdiction-specific ethical requirements

Inputs

Collect the required inputs before generating; ask for any that are missing.

Required: research question / hypothesis (specific, testable); study type (retrospective cohort, prospective cohort, cross-sectional, RCT, diagnostic accuracy, case-control, case series); target population (who, where, when); primary outcome; secondary outcomes (if any).
Optional: key references (DOIs or search terms); institution name (header and ethics guidance); regulatory context -- Korea (PIPA), US (HIPAA/Common Rule), EU (GDPR), other.

Use prior skill outputs directly when they exist; when they do not, ask the user or call the skill:

design-study: design recommendations, analysis unit, comparator design, validation strategy.
calc-sample-size: protocol/sample_size_justification.md (canonical IRB-ready prose) and protocol/sample_size_calc.{R,py} (reproducible code).
search-lit: Background references with verified citations.
define-variables: variable_operationalization.md -- literature-grounded definitions, cutoffs, DB-variable mappings for the Methods section. Precondition: if the study is observational and no operationalization artifact exists, call /define-variables before drafting Methods. Do not invent phenotype/cutoff definitions from the data dictionary inside this skill.

Protocol Structure -- 10 Sections

Core Sections (Fully Generated)

Section 1: Background and Rationale (400-600 words)

Full paragraphs, no bullet points, flowing from clinical context (disease burden, current

Track record

Benchmark

6 / 18

test cases passed, vs 2 / 18 without

Adoption

Users: 0
Times used: 0

Safety and source details

SkillSafety

✓ 3 of 3 passed

Code scan15 of 15 checks clean

0 critical · 0 warning · 0 info — scanner v5 — 10/3/2026

View 15 scanner checks

live_secretremote_code_executionreverse_shelldata_exfiltrationssrf_metadataagent_snoopinganti_refusaldestructive_commandsobfuscated_unicodeunicode_homoglyphopaque_blobsuspicious_urltool_overreachtrigger_abuseprompt_injection_phrasing

IntentSafe

LLM judge · does the skill do what it claims

ContentClean

LLM judge · policy & harmful-content review

Source

Imported

aperivue / medsci-skills

skills/write-protocol/SKILL.md

View on GitHub ↗

License MITImported 10/3/2026Version v3Bundled 2 files

More skills from the same source and nearby in the registry, with their headline benchmark lift.

Ready to supercharge your AI agents?

See it in action

What's inside

Write-Protocol Skill

Inputs

Protocol Structure -- 10 Sections

Core Sections (Fully Generated)

Section 1: Background and Rationale (400-600 words)

Track record

Benchmark

Adoption

SkillSafety

Source

Related skills