Ready to supercharge your AI agents?

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

Get Started Free →

Assumption Challenge · yogsoth-ai | DecimalAI

yogsoth-ai/convergence-assumption-challenge

NewSkillScoreMaintained

Construct the strongest counter-argument against a specific assumption and propose alternatives.

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-convergence-assumption-challenge/SKILL.md

decimalai skills pull yogsoth-ai-convergence-assumption-challenge --out .claude/skills/

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

curl -fsSL --create-dirs -o .claude/skills/yogsoth-ai-convergence-assumption-challenge/SKILL.md https://app.decimal.ai/s/yogsoth-ai-convergence-assumption-challenge/SKILL.md

Raw file: /s/yogsoth-ai-convergence-assumption-challenge/SKILL.md ↗ · machine-readable summary at /s/yogsoth-ai-convergence-assumption-challenge.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 7d ago

—+20% tokens0% turns14% of whole cases pass with skill (3/22)

Test case	Without → With	Effect	Δ tokens	Δ turns
case-06	✓→✓	＝ Same ✓	6%	0%
case-07	✓→✓	＝ Same ✓	93%	0%
case-08	✓→✓	＝ Same ✓	66%	0%
case-01	✗→✗	＝ Same ✗	-58%	0%
case-02	✗→✗	＝ Same ✗	19%	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 — 8/3/2026

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/convergence-assumption-challenge/SKILL.md

View on GitHub ↗

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

Skill contents

SKILL.md

Assumption Challenge

Attacks a specific assumption adversarially — constructing the strongest argument for why it might be wrong, proposing an alternative assumption, and assessing the impact on the overall decision if the assumption fails.

Execution

Spawns a subagent that takes a single assumption and builds the strongest possible case against it.

Why Subagent

Challenge requires adversarial mindset isolated from the assumption-maker
Each assumption deserves focused, dedicated attack
Isolation prevents the challenger from being "too kind" to assumptions

HARD-GATE

Output must include:

Challenge argument with evidence or reasoning
At least 1 alternative assumption (what if the opposite is true?)
Impact assessment if the assumption is wrong
Confidence in the challenge itself (is this a real threat or theoretical?)

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. |

Want to see if yogsoth-ai/convergence-assumption-challenge would activate on your phrasing? Try it before you install — no LLM call, no signup.

Latest benchmark · publisher runVerified

8/22/2026, 08:26 UTC · 22 cases

Measured on 20 comparable cases — below the 22 needed to publish a lift, so this result is shown here but not on the registry card or leaderboard.

Lift

0%

14% vs 14%

Agent turns

0%

1.0 vs 1.0

Tokens

+20%

2,872.9 vs 2,386.2

pass → pass · 3fail → fail · 19

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-01 Our financial model relies on the assumption that customer acquisition costs will decrease by 15% next quarter due to organic referral loops. Please aggressively critique this specific assumption. I want to see a strong case against it, at least one alternative scenario to plan for, an analysis of what happens to our runway if we're wrong, and your confidence assessment on whether this counter-perspective represents a realistic threat.	fail→fail	20,916	17,089	-18%	1	1	0%	3,203	1,354	-58%	0	0	—
case-02 Our product team is building the 2025 growth strategy based on the key premise that enterprise customers will renew at 95% annual net retention once the new reporting module ships. The team is confident and wants a quick check. Run a critique of this premise.	fail→fail	14,902	16,554	+11%	1	1	0%	2,254	2,671	+19%	0	0	—
case-03 Our manufacturing plan for the Apex-5 IoT Gateway assumes chip delivery lead times will remain stable at 6 weeks throughout Q3. Validate this assumption for our executive deck.	fail→fail	14,615	14,391	-2%	1	1	0%	2,115	2,369	+12%	0	0	—
case-04 Engineering assumes the Postgres database migration for the main payments platform will complete within a 2-hour maintenance window on Sunday at 2 AM. Provide a critical review of this assumption.	fail→fail	18,947	18,711	-1%	1	1	0%	2,598	2,959	+14%	0	0	—
case-05 Marketing assumes increasing the monthly subscription price of CloudSync Pro from $19 to $29 will cause less than 5% churn among active users. Evaluate this assumption.	fail→fail	15,207	17,234	+13%	1	1	0%	2,377	3,009	+27%	0	0	—
case-06 We are launching a new consumer mobile app next month. Please generate a standard risk register table listing the top 5 operational risks, their likelihood, impact score, and mitigation strategies.	pass→pass	10,992	10,948	-0%	1	1	0%	1,834	1,948	+6%	0	0	—
case-07 Calculate a 3x3 financial sensitivity matrix showing projected annual revenue for our ecommerce platform across conversion rates of 1.5%, 2.0%, and 2.5% combined with average order values of $40, $50, and $60.	pass→pass	8,958	17,628	+97%	1	1	0%	1,891	3,642	+93%	0	0	—
case-08 According to standard AWS S3 documentation, confirm the maximum single object upload size supported by Amazon S3.	pass→pass	3,280	3,876	+18%	1	1	0%	487	806	+66%	0	0	—
case-09 Our architecture proposal for the RealTime-NLP service relies on the assumption that GPU batching will keep p99 inference latency below 50ms under peak traffic of 10,000 requests per second. Attack this assumption.	fail→fail	19,925	21,135	+6%	1	1	0%	3,101	3,440	+11%	0	0	—
case-10 The trial protocol for Phase II of BioMed-844 assumes patient recruitment across 10 site centers will reach target enrollment within 90 days. Challenge this operational assumption.	fail→fail	15,899	15,049	-5%	1	1	0%	2,357	2,415	+2%	0	0	—
case-11 Our third-party data aggregator assumes the Twitter/X v2 API endpoint will sustain 500 requests per minute without hitting secondary rate limits or HTTP 429 errors. Critique this technical premise.	fail→fail	17,953	18,326	+2%	1	1	0%	2,786	2,910	+4%	0	0	—
case-12 The Q4 workforce plan assumes voluntary attrition among senior DevOps engineers will stay under 2% after the office re-location to downtown. Interrogate this assumption.	fail→fail	14,170	17,205	+21%	1	1	0%	2,022	2,720	+35%	0	0	—
case-13 Our product roadmap assumes the community will build and maintain custom connectors for top 10 CRM tools within 6 months of open-sourcing CoreEngine-v1. Challenge this adoption assumption.	fail→fail	15,357	14,834	-3%	1	1	0%	2,331	2,416	+4%	0	0	—
case-14 The SecOps incident response plan assumes that automated isolation scripts will contain any ransomware outbreak within 15 minutes of detection. Adversarially critique this premise.	fail→fail	18,423	22,084	+20%	1	1	0%	2,545	3,141	+23%	0	0	—
case-15 The financial forecast for expansion into London assumes office lease rates per square foot in Shoreditch will decline by 10% before Q4 contract execution. Challenge this market assumption.	fail→fail	15,022	17,297	+15%	1	1	0%	2,185	2,842	+30%	0	0	—
case-16 The investment memo for Project Helios assumes annual solar panel degradation across the 50MW farm will not exceed 0.5% per year over 20 years. Attack this technical assumption.	fail→fail	19,976	25,021	+25%	1	1	0%	3,026	4,238	+40%	0	0	—
case-17 The holiday fulfillment strategy assumes regional carrier fulfillment times will remain under 48 hours for standard ground shipping during peak December volume. Critically challenge this logistics assumption.	fail→fail	17,243	18,748	+9%	1	1	0%	2,432	2,925	+20%	0	0	—
case-18 Our fintech launch plan assumes FINRA broker-dealer registration approval will be granted within 6 months of application filing. Stress-test this regulatory assumption.	fail→fail	22,713	18,578	-18%	1	1	0%	3,148	2,902	-8%	0	0	—
case-19 FinOps assumes migrating production workloads from Amazon EKS on On-Demand EC2 to Karpenter-managed Spot instances will achieve 60% compute cost reduction without downtime. Challenge this assumption.	fail→fail	21,010	20,795	-1%	1	1	0%	2,888	3,425	+19%	0	0	—
case-20 The growth team assumes removing credit card requirements during the 14-day free trial will double trial-to-paid conversions for SaaS-Flow. Critique this growth assumption.	fail→fail	16,253	15,132	-7%	1	1	0%	2,508	2,543	+1%	0	0	—
case-21 Cloud Architecture assumes cross-region data transfer from Azure East US to AWS us-east-1 via Direct Connect will achieve 10 Gbps sustained throughput without packet loss. Challenge this network performance assumption.	fail→fail	19,222	25,993	+35%	1	1	0%	2,839	4,137	+46%	0	0	—
case-22 The mobile engineering team assumes continuous background location polling will consume less than 3% total daily battery on iOS devices running app version 3.2. Adversarially critique this assumption.	fail→fail	19,642	13,556	-31%	1	1	0%	2,741	1,093	-60%	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, and 20 counted toward the lift figure. The other 2 produced results that are not comparable between the two arms, so they are excluded from the headline rather than averaged into it. The headline lift of 0 percentage points is the difference between those two pass rates over the 20 comparable cases. 2 cases got worse with the skill loaded, and they are 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: 13.6%
Pass rate with the skill: 13.6%
Cases attempted: 22
Cases counted toward lift: 20
Cases that regressed: 2
Skill version measured: v1
Run completed: 8/22/2026, 08:26 UTC

SkillScore blends 4 signalsno signals yet

BenchmarkRun a verified benchmark with positive lift

Live evalNeeds ≥20 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+52% tokens

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

No signals yet — run a benchmark or collect evaluated traces to earn a score.

Verified on gemini-3.6-flash · re-tested 7 days ago

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