Vibe Drops · Week of Nov 2–8, 2026

Best of Week 5

Every week I turn standout creator essays into interactive explorables — tools, not threads. Here are Week 5's 10 drops, each built from its creator's own framework, with their attribution baked in. Click any card to run it.

Tim's Picks

Tim's Week 5 picks

One winner per adoption phase, in phase order — explore, pilot, scale, embed — plus an honorable mention. Tap any row to jump to the card.

1
The Context-Forgetting Simulator8/10
TestMu AI · uniqueness 4, value 4Explore
2
The Eval Predictiveness Check9/10
Edward Grundy · uniqueness 4, value 5Pilot
3
The MCP Schema Rent Check8/10
GetUnblocked · uniqueness 4, value 4Scale
4
The Approval Fatigue Check9/10
Cobus Greyling · uniqueness 5, value 4Embed
★
The OWASP ASI Exposure Audit8/10
OWASP GenAI Security Project · uniqueness 3, value 5 · honorable mentionScale

The OWASP ASI Exposure Audit

Ten questions, one per ASI risk, each answer revealing the OWASP case and the control that closes the gap — verdict ranks all ten risks most-exposed first.

Uniqueness3
Business value5
Open the explorable →

The Pilot-to-Production Check

Eight questions against the 12-item production-readiness checklist and the four phases: Scope, Evaluate, Integrate, Operate.

Uniqueness2
Business value4
Open the explorable →

The Review Queue Stress Test

Six questions on the six-signal trigger taxonomy, calibration, four-tier routing, and the capacity arithmetic.

Uniqueness4
Business value5
Open the explorable →

The Recurring Agent Monitor

Seven questions on the workflow contract, the canonical run record, run-health alerts, and quality-drift thresholds.

Uniqueness1
Business value3
Open the explorable →

The Eval Predictiveness Check

Eight questions on task set, trace, three measurements in order, repetition, and rerun triggers.

Uniqueness4
Business value5
Open the explorable →

The Approval Fatigue Check

Six questions on decision load, forced thought, the hollow-oversight signal, and skill atrophy.

Uniqueness5
Business value4
Open the explorable →

The MCP Schema Rent Check

Seven questions on the context tax, the accuracy curve, platform ceilings, and the Tool Search fix.

Uniqueness4
Business value4
Open the explorable →

The Multi-Agent Reliability Audit

Six-check audit across the four reliability layers: eval gate, prompt versioning, replayable event log, cost per user.

Uniqueness2
Business value3
Open the explorable →

The Prompt-Change Release Builder

Version the five-part decision set, regression-test against fixed past cases, and ship with a named approver — the record is yours to keep.

Uniqueness5
Business value2
Open the explorable →

The Context-Forgetting Simulator

Break-it simulator for Drew Breunig’s four context failure modes: poisoning, distraction, confusion, clash.

Uniqueness4
Business value4
Open the explorable →

Scorecard

Your ratings, ranked by total. Saved on this device.

Top pick per phase

Full ranking

DropPhaseUniqBizTotal