Ready to supercharge your AI agents?

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

Get Started Free →

PICO Formulation SOP · yogsoth-ai | DecimalAI

yogsoth-ai/pico-formulation

SkillScore77Maintained

Construct PICO/PECO framework for the meta-analysis research question

researchby yogsoth-ai

Links this skill to your agents and keeps it current with the author’s updates.

Automatic

Research pack

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

pip install decimalai

2 · Get the skill — writes .claude/skills/yogsoth-ai-pico-formulation/SKILL.md

decimalai skills pull yogsoth-ai-pico-formulation --out .claude/skills/

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

curl -fsSL --create-dirs -o .claude/skills/yogsoth-ai-pico-formulation/SKILL.md https://app.decimal.ai/s/yogsoth-ai-pico-formulation/SKILL.md

Raw file: /s/yogsoth-ai-pico-formulation/SKILL.md ↗ · machine-readable summary at /s/yogsoth-ai-pico-formulation.json ↗

In productionGlobal

how this skill behaves on real traffic · last 30 days

No production data for this skill yet — usage numbers appear here once agents start running it.

Anonymized aggregate across every org using this skill — proof from real production use, not a demo.

Does this skill help? · Benchmark: with vs. withoutsign in to run this on your agents

22 test cases from the publisher’s eval.yaml · each run twice on gemini-3.6-flash — with the skill injected, then without · last run 14d ago

Yes — +18% lift+15% tokens0% turns77% of whole cases pass with skill (17/22)

Test case	Without → With	Effect	Δ tokens	Δ turns
case-18	✗→✓	▲ Improved	46%	0%
case-03	✗→✓	▲ Improved	5%	0%
case-07	✗→✓	▲ Improved	65%	0%
case-13	✗→✓	▲ Improved	6%	0%
case-22	✗→✓	▲ Improved	-6%	0%

Showing 5 of 22 cases

SkillSafety

✓ 3 of 3 passed

Code scan15 of 15 checks clean

0 critical · 0 warning · 0 info — scanner v4 — 8/10/2026

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

yogsoth-ai / de-anthropocentric-research-engine

skills/pico-formulation/SKILL.md

View on GitHub ↗

License Apache-2.0Imported 8/10/2026Version v1

Skill contents

SKILL.md

PICO Formulation SOP

Construct a structured PICO (Population, Intervention, Comparator, Outcome) or PECO (Population, Exposure, Comparator, Outcome) framework from a research question to guide systematic meta-analysis planning.

When to Use

First step in any meta-analysis strategy
When the research question needs formal structuring
Before defining inclusion/exclusion criteria

Input

research_question: The research question to structure
domain: The research domain (clinical, computational, social science, etc.)

Output

A complete PICO/PECO framework with operationalized definitions for each component, suitable for driving search strategy and inclusion criteria.

Available SOPs

Optional, no fixed order; the final leaf is always a sop.

| SOP | When to use | | --- | --- | | spawn-agent | Spawn a customized CC subagent with full MCP tool access. Used by SOPs that declare execution: subagent. |

Illustrative Want to see what yogsoth-ai/pico-formulation would put in front of your model? Preview the rendered body and check its trigger phrases — no LLM call, no signup.

Latest benchmark · publisher runVerified

9/7/2026, 09:01 UTC · 22 cases

Lift

+18%

77% vs 59%

Agent turns

0%

1.0 vs 1.0

Tokens

+15%

2,542.23 vs 2,207.59

fail → pass · 5pass → pass · 12fail → fail · 4pass → fail · 1

Per-case results (22)

Case	Status	Duration (ms)			Turns			Tokens			Tool calls
Case	Status	Without	With	Δ	Without	With	Δ	Without	With	Δ	Without	With	Δ
case-19 Evaluating alternating pressure mattresses vs standard foam mattresses for pressure ulcer prevention in bedridden ICU patients. I plan to group both mattress types into a single Intervention group. Provide the framework breakdown.	pass→pass	24,500	24,176	-1%	1	1	0%	3,125	3,661	+17%	0	0	—
case-01 I am planning a systematic review comparing metformin versus lifestyle interventions for glycemic control in adult patients with prediabetes. I want to merge metformin and lifestyle into a single Intervention element called 'Active Interventions' and compare against placebo. Provide the structured PICO breakdown.	fail→fail	15,324	15,784	+3%	1	1	0%	1,779	2,214	+24%	0	0	—
case-02 We are designing a systematic review on ambient fine particulate matter (PM2.5) air pollution and cardiovascular mortality in urban populations. I plan to use a standard clinical PICO framework with PM2.5 classified as an Intervention arm. Provide the framework breakdown.	pass→pass	18,189	20,078	+10%	1	1	0%	2,382	2,895	+22%	0	0	—
case-18 We are studying prenatal lead exposure and neurodevelopmental delays in toddlers. I plan to frame prenatal lead exposure as an Intervention in a standard PICO framework. Provide the framework breakdown.	fail→pass	17,090	19,405	+14%	1	1	0%	1,893	2,755	+46%	0	0	—
case-03 We are evaluating a novel graph neural network architecture against traditional convolutional neural networks for molecular property prediction on the QM9 dataset. I intend to set the QM9 dataset as the Intervention component. Provide the framework breakdown.	fail→pass	20,314	18,619	-8%	1	1	0%	2,536	2,652	+5%	0	0	—
case-04 I want to structure a meta-analysis on universal basic income (UBI) trials versus traditional conditional welfare programs for reducing poverty among low-income households. I plan to classify UBI as an observational Exposure and omit a comparator. Provide the framework breakdown.	fail→fail	25,510	21,433	-16%	1	1	0%	3,390	2,794	-18%	0	0	—
case-05 A researcher wants to evaluate flipped classroom teaching methods versus standard lectures for undergraduate computer science students regarding exam scores. They want to set 'undergraduate computer science' as the Intervention component. Provide the framework breakdown.	pass→pass	14,235	18,018	+27%	1	1	0%	1,476	2,479	+68%	0	0	—
case-06 We are assessing night-shift work duration and breast cancer risk among healthcare workers. A team member wants to frame night-shift work as a clinical Intervention in a PICO framework. Provide the framework breakdown.	pass→pass	14,320	17,243	+20%	1	1	0%	1,663	2,350	+41%	0	0	—
case-07 For elderly patients with severe lumbar spinal stenosis, we want to compare surgical decompression with physical therapy for pain reduction. I plan to combine physical therapy and surgery as a single Intervention arm against no treatment. Provide the framework breakdown.	fail→pass	17,007	23,996	+41%	1	1	0%	2,051	3,381	+65%	0	0	—
case-08 We are structuring a systematic review evaluating rapid antigen detection tests against RT-PCR for detecting influenza A in pediatric emergency room patients. I plan to use a standard therapeutic PICO framework with RT-PCR as the Intervention. Provide the framework breakdown.	pass→pass	18,144	16,859	-7%	1	1	0%	2,342	2,655	+13%	0	0	—
case-09 A trial investigates mobile app-based cognitive behavioral therapy (CBT) versus face-to-face CBT for depression severity in young adults. I want to leave the Comparator component blank since both groups receive CBT. Provide the framework breakdown.	pass→fail	15,144	14,823	-2%	1	1	0%	1,773	1,788	+1%	0	0	—
case-10 We are studying ultra-processed food consumption and incidence of type 2 diabetes in prospective cohort studies. I plan to structure ultra-processed food intake as a clinical Intervention in a PICO framework. Provide the framework breakdown.	fail→fail	19,448	17,318	-11%	1	1	0%	2,418	2,216	-8%	0	0	—
case-11 A review aims to evaluate prophylactic probiotic supplementation versus placebo for acute diarrhea incidence in shelter dogs. I plan to list 'acute diarrhea' under Population and leave Outcome blank. Provide the framework breakdown.	pass→pass	12,508	13,109	+5%	1	1	0%	1,314	1,633	+24%	0	0	—
case-12 We are planning a meta-analysis on community mask mandates during viral epidemics to reduce transmission rates. I intend to place 'community mask mandates' as the Population element in the framework. Provide the framework breakdown.	pass→pass	18,582	17,630	-5%	1	1	0%	2,222	2,450	+10%	0	0	—
case-13 Comparing high-dose vs standard-dose methotrexate in pediatric acute lymphoblastic leukemia for 5-year event-free survival. I plan to divide the Population into two separate populations: 'high-dose pediatric patients' and 'standard-dose pediatric patients'. Provide the framework breakdown.	fail→pass	20,757	20,834	+0%	1	1	0%	2,985	3,158	+6%	0	0	—
case-14 Evaluating presence of the APOE e4 allele versus non-carriers for Alzheimer's disease onset age. I plan to structure this using PICO with APOE e4 allele as a therapeutic Intervention. Provide the framework breakdown.	pass→pass	17,238	18,256	+6%	1	1	0%	2,148	2,528	+18%	0	0	—
case-15 A review investigates mindfulness-based stress reduction (MBSR) versus active relaxation training for burnout reduction in healthcare workers. I plan to put 'burnout' into Population as 'burned-out workers' and leave Outcome empty. Provide the framework breakdown.	pass→pass	15,660	19,350	+24%	1	1	0%	1,752	2,498	+43%	0	0	—
case-16 We want to structure a review comparing cost-effectiveness of robotic-assisted vs laparoscopic prostatectomy in prostate cancer patients. I plan to exclude financial costs from Outcome and list only clinical outcomes. Provide the framework breakdown.	fail→fail	20,104	17,325	-14%	1	1	0%	2,797	2,477	-11%	0	0	—
case-17 Comparing rough-surface vs smooth-surface dental implants for 10-year failure rates in adult periodontal patients. I plan to list 'dental implants' as the Population element. Provide the framework breakdown.	pass→pass	14,259	19,847	+39%	1	1	0%	1,658	2,751	+66%	0	0	—
case-20 I have already defined my PICO framework components (Population: adult diabetics, Intervention: SGLT2 inhibitors, Comparator: DPP-4 inhibitors, Outcome: cardiovascular mortality). Please write the exact PubMed MeSH and boolean search query string with appropriate field tags.	pass→pass	19,426	18,124	-7%	1	1	0%	3,122	2,936	-6%	0	0	—
case-21 We have finished screening 1,200 records from PubMed and Embase, removing 300 duplicates, excluding 700 at title/abstract stage, and assessing 200 full texts. Generate the markdown summary table representing a PRISMA flow diagram for our manuscript.	pass→pass	13,747	12,674	-8%	1	1	0%	1,713	1,753	+2%	0	0	—
case-22 Here is a summary of a randomized controlled trial comparing drug X to placebo, noting blinding was broken for outcome assessors. Assess the risk of bias for outcome measurement according to Cochrane RoB 2 tool domain 4 criteria.	fail→pass	17,480	13,750	-21%	1	1	0%	2,028	1,905	-6%	0	0	—

These cases were also screened by the skill’s intent / safety pass.See the Trust review ↗Reproduce this benchmark in Colab ↗

How this skill was measured

DecimalAI ran this skill against gemini-3.6-flash twice over the same eval suite — once with the skill loaded and once without — and compared the two runs case by case. 22 cases were attempted. The headline lift of +18 percentage points is the difference between those two pass rates over the 22 comparable cases. 1 case got worse with the skill loaded, and it is included in that figure.

Model measured: gemini-3.6-flash
Graded by: gemini-3.6-flash
Grading method: judged
Verification: verified
Pass rate without the skill: 59.1%
Pass rate with the skill: 77.3%
Cases attempted: 22
Cases counted toward lift: 22
Cases that regressed: 1
Skill version measured: v1
Run completed: 9/7/2026, 09:01 UTC

Measured example — same prompt, with vs. without the skill · from the benchmark run · case case-18

Prompt

We are studying prenatal lead exposure and neurodevelopmental delays in toddlers. I plan to frame prenatal lead exposure as an Intervention in a standard PICO framework. Provide the framework breakdown.

Without the skill loaded, the model failed this case. With it loaded, the same prompt on the same model passed. This is one improved case from the latest verified run; every case, including any that regressed, is in the table above.

Show both answers

SkillScore blends 4 signals1 of 4 complete

BenchmarkPass rate WITH skill on a verified run77

Live evalNeeds ≥10 evaluated traces in 30 days

AI ratingNeeds ≥10 rated traces

AdoptionNeeds ≥10 serves across ≥2 orgs in 30 days

Efficiency · separate axis (at equal correctness)+0% turns+23% tokens

Shown for transparency; not credited unless a verified run shows no correctness regression.

Backed by 1 of 4 signals — each added signal makes the score more robust.

Verified on gemini-3.6-flash · re-tested 14 days ago

Other measured skills in the registry, with their headline benchmark lift.