bdk

Plugin bdk from marketplace ai-baloney-detection-kit in deanpeters/ai-baloney-detection-kit. Commands read from the repo's manifest — 21 days ago.

Skills (13)

• bdk-context-pantry

  Find decision-critical context, contradictions, and gaps in an AI idea - workflow, ownership, and both sides' economics. Use when the customer, outcome, workflow, economics, or constraints are unclear; a deep-dive companion to bdk-relevance in the Confusion tier.

• bdk-detector

  Run one AI idea or a whole pile (a roadmap, vendor pitches) through the seven-step Baloney Funnel - outcome, receipts, labels, sell-by, relevance, taste test, pivot-punt-pursue - and report what was sifted out where, and what earned Pivot, Punt, or Pursue.

• bdk-four-risks

  Read the four-risk grid (valuable, usable, viable, feasible) for a named investment and show which square decides. Use for a deep dive on the four risks; the grid behind bdk-pivot-punt-pursue in the Considered tier.

• bdk-labels

  Separate fact, inference, assumption, and best guess in the evidence behind an AI outcome, claim, or product document, and show how much of it is which. Use when polished wording or repeated claims hide uncertainty; step 3 of the Baloney Funnel, in the Confusion tier.

• bdk-outcomes

  Shape an AI pitch, claim, or idea into a measurable outcome statement (direction, metric, target, timeframe, beneficiary, context), or sift it out as output-only or vanity. Use first, before checking any evidence; step 1 of the Baloney Funnel, in the Chaos tier.

• bdk-pivot-punt-pursue

  Recommend Pivot, Punt, or Pursue for an AI bet, and say why, from its outcome, its relevant, labeled, fresh receipts, and the tiny tests run on the four risks (valuable, usable, viable, feasible). Use when it's time to decide; step 7 of the Baloney Funnel, the Considered tier.

• bdk-receipts

  Trace AI claims, and the evidence behind an outcome's metric and direction, to original or authoritative sources and assess support and independence. Use when citations, provenance, or source-of-truth records need checking; step 2 of the Baloney Funnel, in the Chaos tier.

• bdk-relevance

  Judge whether the evidence behind an AI outcome applies to us, directly, through an adjacent product or segment, or by analog, using our pantry of product, user, business, and market data. Use after evidence is labeled and dated; step 5 of the Baloney Funnel, in the Confusion tier.

• bdk-sell-by

  Assess whether evidence is fresh enough for a specific AI decision and set re-check dates. Use for changing prices, models, policies, benchmarks, markets, or experiments; step 4 of the Baloney Funnel, in the Confusion tier.

• bdk-skill-sorter

  Sort a product role's work onto the Sort grid (busywork, beige, theater, truth), give every item its move (pass-off, pivot, punt, pursue), then say what it means for getting hired, fired, and promoted, and what to do Monday. Use with a job description, a role, a calendar, or a list of what someone d

• bdk-taste-test

  Design the tiniest tests that return the most brutal truth on whether an AI bet is valuable, usable, viable, and feasible, or review supplied results. Use before a bet advances beyond plausible output or a working demo; step 6 of the Baloney Funnel, the Crapola tier.

• bdk-util-slice-the-pitch

  Pull the actual, checkable claims out of a deck, demo, press release, roadmap, or vendor page so they can be poured into the Baloney Funnel. Use when you have a document, not a list of claims.

• bdk-value-or-theater

  Put an AI initiative, vendor pitch, roadmap item, or LinkedIn hot take through five questions and rule it value, theater, or not yet. Use on any AI pitch or roadmap item for a fast verdict; it is the quick check shown live in the ProductFest 2026 keynote.

Similar plugins