Ready to supercharge your AI agents?

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

Adaptive Pair Selection · yogsoth-ai | DecimalAI

researchby yogsoth-ai

Adaptive Pair Selection

yogsoth-ai/adaptive-pair-selection

Source yogsoth-ai/de-anthropocentric-research-engine ↗License Apache-2.0

Iteratively select maximally informative pairs, execute comparisons, update ratings, and check convergence until ranking stabilizes.

SkillScore91Maintained

Automatic

AI Engineer pack

Last verified 8/22/2026

What it changes

Passed 20 of 22 test cases with the skill, and 15 of 22 without it.

Without skill68%

With skill91%

Checks it passed that the model alone missed

The response lists judgment, current_ratings, and method as the required input parameters.
The response specifies that convergence-check returns converged status and stability_score.
The response requires providing confidence intervals for each position in the global ranking.

and 2 more in the full report

5 fixed15 passed either way2 still failing0 broke

Pass-rate lift: +23 pp
Tokens: -18%
Agent turns: 0%
Response time: -47%

Model gemini-3.6-flashMeasured 8/22/2026v120 of 22 cases passed with skill

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

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/yogsoth-ai-adaptive-pair-selection/SKILL.md https://app.decimal.ai/s/yogsoth-ai-adaptive-pair-selection/SKILL.md

Want it pinned and updatable? pip install "decimalai>=0.15", then decimalai skills add yogsoth-ai-adaptive-pair-selection, 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/yogsoth-ai-adaptive-pair-selection/ and pins it in .decimal/skills.lock

decimalai skills add yogsoth-ai-adaptive-pair-selection

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

curl -fsSL --create-dirs -o .claude/skills/yogsoth-ai-adaptive-pair-selection/SKILL.md https://app.decimal.ai/s/yogsoth-ai-adaptive-pair-selection/SKILL.md

Raw file: /s/yogsoth-ai-adaptive-pair-selection/SKILL.md ↗ · machine-readable summary at /s/yogsoth-ai-adaptive-pair-selection.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

When running the actual head-to-head comparison between Candidate 12 and Candidate 4, which sub-component SOP is responsible for producing the raw judgment along with confidence level and rationale?

What was checked, without and with the skill
Checked	Without skill	With skill
The response specifies comparison-executor as the SOP responsible for producing the head-to-head judgment.	Missed	Passed

case-07 · gemini-3.6-flash · 8/22/2026 · answers are capped and redacted

What's inside

The instructions your model reads when the skill loads.

SKILL.md51 linessingle file

Adaptive Pair Selection

Select the next comparison pair by information gain, execute the comparison, update ratings, and check for convergence. Repeats until the ranking stabilizes or the comparison budget is exhausted.

Stages

Select — pair-selector identifies the pair whose comparison would most reduce uncertainty
Compare — comparison-executor produces a judgment with confidence and reasoning
Update — rating-update incorporates the new judgment into the rating model
Check — convergence-check determines if ranking has stabilized

Loop stages 1-4 until convergence or budget exhaustion.

Available SOPs

Stage	SOP	Input	Output
Select	pair-selector	current_ratings, comparison_history	next_pairs[]
Compare	comparison-executor	pair, context	judgment
Update	rating-update	judgment, current_ratings, method	updated_ratings
Check	convergence-check	rating_history	converged, stability_score

Execution Guidance

Start with high-uncertainty pairs (largest sigma or most uncertain boundary)
For small N: may complete all pairs in first pass, then focus on inconsistencies
For large N: prioritize pairs near rank boundaries (positions k and k+1)
Track comparison count against budget; exit gracefully if budget hit
Pass full rating_history to convergence-check (not just latest snapshot)

Minimum Yield

Global ranking + confidence intervals + convergence curve
Global ranking with confidence intervals for each position
Convergence curve showing stability score over iterations
Comparison log with all judgments made

Available SOPs

Track record

Benchmark

20 / 22

test cases passed, vs 15 / 22 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 v4 — 8/3/2026

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

yogsoth-ai / de-anthropocentric-research-engine

skills/adaptive-pair-selection/SKILL.md

View on GitHub ↗

License Apache-2.0Imported 8/3/2026Version v1

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

Adaptive Pair Selection

Stages

Available SOPs

Execution Guidance

Minimum Yield

Available SOPs

Track record

Benchmark

Adoption

SkillSafety

Source

Related skills