One row per video; its state is whoever acts next. A chain of questions is one row; its follow-up questions have none. Click a header to sort; filter by text, topic or state. The state comes from the committed artifacts: readiness from stories.json, then the first production stage whose file is missing or not current (re-running the library on today's inputs must reproduce the file, the test lock applies). The strip in each row is the production stages: what has run (filled), the stage it is at, and what is still to come; red means /produce stopped and a person must act. Only production.json is hand-written (queued, parked), because no artifact records those.
16 documents, 15 with verdicts, 54/385 claims decided by a study; 1 documents without a recorded verdict (scripts/briefs/sources.md) topics/ai-coding/sources/
174 listed, 172 in the store; 2 not fetched (npm run watcher -- fetch topics/get-rich-quick/videos.txt --store topics/get-rich-quick/store) topics/get-rich-quick/videos.txt
22 documents, 20 with verdicts, 59/708 claims decided by a study; 2 documents without a recorded verdict (scripts/briefs/sources.md) topics/get-rich-quick/sources/
Find studies or public data that would decide this gap in topic heat-pumps, story conflict:heat-pumps-need-home-well-insulated:
Reviewer: the evidence is NREL's model (envelope upgrades cut bill increases but add about $9,000, helping mainly in cold climates), RMI and LBNL modelling of a few archetypes, and small Minnesota field data, none above 0.45 and none measuring heat pump bills in leaky against insulated homes; a measured comparison of heat pump running cost and payback with and without envelope work, by climate, would unblock it
Save each document to topics/heat-pumps/sources/ as text (pdftotext -layout for PDFs), then follow scripts/briefs/sources.md for the new documents against every claim and record with npm run record-sources -- heat-pumps topics/heat-pumps/work/sv_*.json.
Then run npm run report and npm run stories -- discover heat-pumps. If nothing decides it, say so and stop.
Reviewer: the evidence is NREL's model (envelope upgrades cut bill increases but add about $9,000, helping mainly in cold climates), RMI and LBNL modelling of a few archetypes, and small Minnesota field data, none above 0.45 and none measuring heat pump bills in leaky against insulated homes; a measured comparison of heat pump running cost and payback with and without envelope work, by climate, would unblock it
Find studies or public data that would decide this gap in topic organic-food, story conflict:pesticide-residues-conventional-produce-too-low:
Reviewer: The corpus answers this only indirectly: a 2011 risk model of mean exposure to the ten most detected pesticides on 2010 Dirty Dozen crops, a USDA summary of tolerance compliance that measures no health outcome, and a review's summary of organophosphate birth cohorts mostly from farmworker and inner-city settings. Unblocking it needs a direct study of health outcomes against dietary residue exposure in the general population (a cohort linking measured dietary or urinary residues to outcomes, or a current cumulative dietary risk assessment that includes high-end consumers and pregnancy).
Save each document to topics/organic-food/sources/ as text (pdftotext -layout for PDFs), then follow scripts/briefs/sources.md for the new documents against every claim and record with npm run record-sources -- organic-food topics/organic-food/work/sv_*.json.
Then run npm run report and npm run stories -- discover organic-food. If nothing decides it, say so and stop.
Reviewer: The corpus answers this only indirectly: a 2011 risk model of mean exposure to the ten most detected pesticides on 2010 Dirty Dozen crops, a USDA summary of tolerance compliance that measures no health outcome, and a review's summary of organophosphate birth cohorts mostly from farmworker and inner-city settings. Unblocking it needs a direct study of health outcomes against dietary residue exposure in the general population (a cohort linking measured dietary or urinary residues to outcomes, or a current cumulative dietary risk assessment that includes high-end consumers and pregnancy).
How often is food sold as organic not really organic?
0.73
0.82
0.60
5
Needs research
Next: decision:can-you-trust-the-label
Give an agent this prompt
Find studies or public data that would decide this gap in topic organic-food, story decision:can-you-trust-the-label:
Reviewer: Every source on this question is the agency describing its own rules or its auditor describing control gaps: documented fraud cases exist, certifiers must test only 5% of operations a year, and the 2017 audit found weak import checks, but nothing in the corpus measures how much organic food is mislabelled or whether the 2023 enforcement rule reduced it. Unblocking it needs a measured rate: residue or isotope testing of retail organic samples, NOP enforcement and residue-testing results over time, or an independent analysis of import volumes against certified production before and after the rule.
Save each document to topics/organic-food/sources/ as text (pdftotext -layout for PDFs), then follow scripts/briefs/sources.md for the new documents against every claim and record with npm run record-sources -- organic-food topics/organic-food/work/sv_*.json.
Then run npm run report and npm run stories -- discover organic-food. If nothing decides it, say so and stop.
Reviewer: Every source on this question is the agency describing its own rules or its auditor describing control gaps: documented fraud cases exist, certifiers must test only 5% of operations a year, and the 2017 audit found weak import checks, but nothing in the corpus measures how much organic food is mislabelled or whether the 2023 enforcement rule reduced it. Unblocking it needs a measured rate: residue or isotope testing of retail organic samples, NOP enforcement and residue-testing results over time, or an independent analysis of import volumes against certified production before and after the rule.
Find studies or public data that would decide this gap in topic organic-food, story conflict:dirty-dozen-list-reliable-guide-which:
Reviewer: The only critique in the corpus (Winter and Katz 2011) assessed the 2010 list and the old six-indicator method; EWG changed its method in 2025 to add a toxicity factor. Unblocking it needs a current, independent assessment of the 2025-2026 list that weights residues by how much of each food people eat and compares the exposure to health thresholds, showing whether the ranking separates high-risk from low-risk produce.
Save each document to topics/organic-food/sources/ as text (pdftotext -layout for PDFs), then follow scripts/briefs/sources.md for the new documents against every claim and record with npm run record-sources -- organic-food topics/organic-food/work/sv_*.json.
Then run npm run report and npm run stories -- discover organic-food. If nothing decides it, say so and stop.
Reviewer: The only critique in the corpus (Winter and Katz 2011) assessed the 2010 list and the old six-indicator method; EWG changed its method in 2025 to add a toxicity factor. Unblocking it needs a current, independent assessment of the 2025-2026 list that weights residues by how much of each food people eat and compares the exposure to health thresholds, showing whether the ranking separates high-risk from low-risk produce.
Find studies or public data that would decide this gap in topic heat-pumps, story conflict:real-world-heat-pump-efficiency-lower:
Reviewer: the field studies split (Cadmus measured below rating, CEE within error at one rating point, NREL matched at 7 of 12 sites) and the seasonal COPs measured (NREL 12 sites averaging 2.5, CEE three Minnesota sites at 2.5-2.8) are all small cold-climate samples; a metered study of many installs across climates, reporting seasonal COP against rated HSPF2, would unblock it
Save each document to topics/heat-pumps/sources/ as text (pdftotext -layout for PDFs), then follow scripts/briefs/sources.md for the new documents against every claim and record with npm run record-sources -- heat-pumps topics/heat-pumps/work/sv_*.json.
Then run npm run report and npm run stories -- discover heat-pumps. If nothing decides it, say so and stop.
Reviewer: the field studies split (Cadmus measured below rating, CEE within error at one rating point, NREL matched at 7 of 12 sites) and the seasonal COPs measured (NREL 12 sites averaging 2.5, CEE three Minnesota sites at 2.5-2.8) are all small cold-climate samples; a metered study of many installs across climates, reporting seasonal COP against rated HSPF2, would unblock it
Find studies or public data that would decide this gap in topic organic-food, story decision:glyphosate:
Reviewer: The harm question rests on one regulator's 2017 dietary assessment, whose human-health decision was vacated in 2022 and is being redone, and the exposure question on a review noting that organic diets cut urinary glyphosate in a few studies, with no comparison to other ways of lowering it. Unblocking it needs an independent dietary risk assessment or biomonitoring study for glyphosate at general-population levels (or EPA's re-evaluation once published) and residue data on how often organic and conventional foods carry glyphosate.
Save each document to topics/organic-food/sources/ as text (pdftotext -layout for PDFs), then follow scripts/briefs/sources.md for the new documents against every claim and record with npm run record-sources -- organic-food topics/organic-food/work/sv_*.json.
Then run npm run report and npm run stories -- discover organic-food. If nothing decides it, say so and stop.
Reviewer: The harm question rests on one regulator's 2017 dietary assessment, whose human-health decision was vacated in 2022 and is being redone, and the exposure question on a review noting that organic diets cut urinary glyphosate in a few studies, with no comparison to other ways of lowering it. Unblocking it needs an independent dietary risk assessment or biomonitoring study for glyphosate at general-population levels (or EPA's re-evaluation once published) and residue data on how often organic and conventional foods carry glyphosate.
Find studies or public data that would decide this gap in topic heat-pumps, story conflict:contractors-routinely-oversize-heating-cooling-equipment:
Reviewer: the only evidence is a dozen NREL field sites, mostly undersized; a large audit comparing installed capacity with Manual J loads would unblock it
Save each document to topics/heat-pumps/sources/ as text (pdftotext -layout for PDFs), then follow scripts/briefs/sources.md for the new documents against every claim and record with npm run record-sources -- heat-pumps topics/heat-pumps/work/sv_*.json.
Then run npm run report and npm run stories -- discover heat-pumps. If nothing decides it, say so and stop.
Reviewer: the only evidence is a dozen NREL field sites, mostly undersized; a large audit comparing installed capacity with Manual J loads would unblock it
Find studies or public data that would decide this gap in topic get-rich-quick, story contradicted:network-marketing-success-learnable-skill-rather:
Reviewer: Only the UK Amway judgment quoted by Taylor bears on skill versus position, and it is one company; the talent-versus-luck simulation is generic. A study relating MLM earnings to join date or training would settle it.
Save each document to topics/get-rich-quick/sources/ as text (pdftotext -layout for PDFs), then follow scripts/briefs/sources.md for the new documents against every claim and record with npm run record-sources -- get-rich-quick topics/get-rich-quick/work/sv_*.json.
Then run npm run report and npm run stories -- discover get-rich-quick. If nothing decides it, say so and stop.
Reviewer: Only the UK Amway judgment quoted by Taylor bears on skill versus position, and it is one company; the talent-versus-luck simulation is generic. A study relating MLM earnings to join date or training would settle it.
Find studies or public data that would decide this gap in topic get-rich-quick, story surprise:commission-free-brokers-sell-order-flow:
Reviewer: The central sentence is that brokers sell order flow, but the only cited passage (Bryzgalova L123-L136) shows the masked spread cost, not the payment for order flow. A verdict citing the paper's own PFOF lines (bryzgalova-retail-options.txt L46-L56, brokerages' PFOF in stocks and options, Q2 2021) or the cost totals (L171 or L956, $4.13bn of spread costs against about $800m of commissions) would unblock it;
Save each document to topics/get-rich-quick/sources/ as text (pdftotext -layout for PDFs), then follow scripts/briefs/sources.md for the new documents against every claim and record with npm run record-sources -- get-rich-quick topics/get-rich-quick/work/sv_*.json.
Then run npm run report and npm run stories -- discover get-rich-quick. If nothing decides it, say so and stop.
2. Do
chain it after decision:trading-and-betting, whose net-of-fees losses raise the question whether zero-commission apps removed the cost.
Reviewer: The central sentence is that brokers sell order flow, but the only cited passage (Bryzgalova L123-L136) shows the masked spread cost, not the payment for order flow. A verdict citing the paper's own PFOF lines (bryzgalova-retail-options.txt L46-L56, brokerages' PFOF in stocks and options, Q2 2021) or the cost totals (L171 or L956, $4.13bn of spread costs against about $800m of commissions) would unblock it; then chain it after decision:trading-and-betting, whose net-of-fees losses raise the question whether zero-commission apps removed the cost.
Find studies or public data that would decide this gap in topic ai-coding, story conflict:generated-code-drifts-codebase-established-conventions:
Reviewer: Only qualitative reports from 16 METR developers and a note on context consistency in the commit study; no measured drift rate against a human baseline.
Save each document to topics/ai-coding/sources/ as text (pdftotext -layout for PDFs), then follow scripts/briefs/sources.md for the new documents against every claim and record with npm run record-sources -- ai-coding topics/ai-coding/work/sv_*.json.
Then run npm run report and npm run stories -- discover ai-coding. If nothing decides it, say so and stop.
Reviewer: Only qualitative reports from 16 METR developers and a note on context consistency in the commit study; no measured drift rate against a human baseline.
When the Iran conflict ends, will fuel prices come back down?
0.61
0.69
0.50
3
Needs research
Next: decision:the-iran-war-and-energy
Give an agent this prompt
Find studies or public data that would decide this gap in topic inflation-2026, story decision:the-iran-war-and-energy:
Reviewer: The two central claims are forecasts (gas returning to its pre-war level, oil staying above it) with no study, and the only evidence on the up-fast-down-slow pattern is two episodes in the weekly EIA series (partly, 0.5). It needs a pass-through study of how quickly US retail gasoline and diesel follow crude down after a supply shock, or post-war price data once the conflict ends, and an energy-market analysis of the risk premium in crude after Gulf supply disruptions.
Save each document to topics/inflation-2026/sources/ as text (pdftotext -layout for PDFs), then follow scripts/briefs/sources.md for the new documents against every claim and record with npm run record-sources -- inflation-2026 topics/inflation-2026/work/sv_*.json.
Then run npm run report and npm run stories -- discover inflation-2026. If nothing decides it, say so and stop.
Reviewer: The two central claims are forecasts (gas returning to its pre-war level, oil staying above it) with no study, and the only evidence on the up-fast-down-slow pattern is two episodes in the weekly EIA series (partly, 0.5). It needs a pass-through study of how quickly US retail gasoline and diesel follow crude down after a supply shock, or post-war price data once the conflict ends, and an energy-market analysis of the risk premium in crude after Gulf supply disruptions.
Find studies or public data that would decide this gap in topic ai-coding, story conflict:writing-code-never-main-bottleneck-software:
Reviewer: Google's migration report and the Microsoft field trial each note that coding is only part of the job, but neither measures end-to-end delivery time by phase; the claim needs a value-stream measurement.
Save each document to topics/ai-coding/sources/ as text (pdftotext -layout for PDFs), then follow scripts/briefs/sources.md for the new documents against every claim and record with npm run record-sources -- ai-coding topics/ai-coding/work/sv_*.json.
Then run npm run report and npm run stories -- discover ai-coding. If nothing decides it, say so and stop.
Reviewer: Google's migration report and the Microsoft field trial each note that coding is only part of the job, but neither measures end-to-end delivery time by phase; the claim needs a value-stream measurement.
Find studies or public data that would decide this gap in topic ai-coding, story conflict:specs-code-workflows-ignore-generated-not:
Reviewer: The code-ownership experiment shows authors understand AI-assisted code less and the vibe-coding review only recommends reading code before committing; nothing measures a spec-only workflow (the spec as source of truth, generated code unread) against a reviewed one on defects, incidents or later change cost.
Save each document to topics/ai-coding/sources/ as text (pdftotext -layout for PDFs), then follow scripts/briefs/sources.md for the new documents against every claim and record with npm run record-sources -- ai-coding topics/ai-coding/work/sv_*.json.
Then run npm run report and npm run stories -- discover ai-coding. If nothing decides it, say so and stop.
Reviewer: The code-ownership experiment shows authors understand AI-assisted code less and the vibe-coding review only recommends reading code before committing; nothing measures a spec-only workflow (the spec as source of truth, generated code unread) against a reviewed one on defects, incidents or later change cost.
Find studies or public data that would decide this gap in topic ai-coding, story surprise:vibe-coded-apps-work-prototypes-but:
Reviewer: The review supports the prototype half; whether vibe-coded apps break down in production is mostly unstudied, so the story needs a case series or longitudinal data.
Save each document to topics/ai-coding/sources/ as text (pdftotext -layout for PDFs), then follow scripts/briefs/sources.md for the new documents against every claim and record with npm run record-sources -- ai-coding topics/ai-coding/work/sv_*.json.
Then run npm run report and npm run stories -- discover ai-coding. If nothing decides it, say so and stop.
Reviewer: The review supports the prototype half; whether vibe-coded apps break down in production is mostly unstudied, so the story needs a case series or longitudinal data.
Do small pull requests really get reviewed faster and break less?
0.58
0.75
0.33
3
Needs research
Next: decision:pairing-and-code-review
Give an agent this prompt
Find studies or public data that would decide this gap in topic software-practices, story decision:pairing-and-code-review:
Reviewer: Sadowski et al. show only that small changes get first feedback sooner at Google, not that they merge faster; McIntosh et al. measure review rate, not review size; DORA ties batch size to failures only through approvals. Unblock with review-to-merge latency broken down by change size (for example Rigby and Bird's cross-project review data), a measured link between lines under review and defects found (for example the Cisco review study), and change size against change failure or restore time.
Save each document to topics/software-practices/sources/ as text (pdftotext -layout for PDFs), then follow scripts/briefs/sources.md for the new documents against every claim and record with npm run record-sources -- software-practices topics/software-practices/work/sv_*.json.
Then run npm run report and npm run stories -- discover software-practices. If nothing decides it, say so and stop.
Reviewer: Sadowski et al. show only that small changes get first feedback sooner at Google, not that they merge faster; McIntosh et al. measure review rate, not review size; DORA ties batch size to failures only through approvals. Unblock with review-to-merge latency broken down by change size (for example Rigby and Bird's cross-project review data), a measured link between lines under review and defects found (for example the Cisco review study), and change size against change failure or restore time.
Find studies or public data that would decide this gap in topic ai-coding, story conflict:goal-loop-commands-prompt-agent-until:
Reviewer: Only Airbnb's bespoke retry pipeline (75% of 3.5K test files on the first bulk run, with a retry cap); needs a measured study of generic goal or loop commands, or background agents assigned routine issues, on real backlogs: share of pull requests merged without rework, human review time per merged change, and regressions introduced.
Save each document to topics/ai-coding/sources/ as text (pdftotext -layout for PDFs), then follow scripts/briefs/sources.md for the new documents against every claim and record with npm run record-sources -- ai-coding topics/ai-coding/work/sv_*.json.
Then run npm run report and npm run stories -- discover ai-coding. If nothing decides it, say so and stop.
Reviewer: Only Airbnb's bespoke retry pipeline (75% of 3.5K test files on the first bulk run, with a retry cap); needs a measured study of generic goal or loop commands, or background agents assigned routine issues, on real backlogs: share of pull requests merged without rework, human review time per merged change, and regressions introduced.
Find studies or public data that would decide this gap in topic organic-food, story contradicted:heavy-glyphosate-exposure-raises-risk-non:
Reviewer: The only source is a summary of the EPA and IARC positions: IARC's pooled case-control analysis found an association, the large Agricultural Health Study cohort found none, and IARC rates the human evidence 'limited'. Unblocking it needs the primary epidemiology itself, such as the Agricultural Health Study cohort results and a meta-analysis of highly exposed groups, so the split can be explained by design and exposure level rather than by two agencies' summaries.
Save each document to topics/organic-food/sources/ as text (pdftotext -layout for PDFs), then follow scripts/briefs/sources.md for the new documents against every claim and record with npm run record-sources -- organic-food topics/organic-food/work/sv_*.json.
Then run npm run report and npm run stories -- discover organic-food. If nothing decides it, say so and stop.
Reviewer: The only source is a summary of the EPA and IARC positions: IARC's pooled case-control analysis found an association, the large Agricultural Health Study cohort found none, and IARC rates the human evidence 'limited'. Unblocking it needs the primary epidemiology itself, such as the Agricultural Health Study cohort results and a meta-analysis of highly exposed groups, so the split can be explained by design and exposure level rather than by two agencies' summaries.
Should a new project start as microservices or as a monolith?
0.57
0.75
0.30
4
Needs research
Next: decision:microservices-vs-monolith
Give an agent this prompt
Find studies or public data that would decide this gap in topic software-practices, story decision:microservices-vs-monolith:
Reviewer: The only sources are a review of practitioner reports (Soldani 2018) that lists pains without measuring cost or outcomes, a DORA footnote advising against premature decomposition with no data, and coordination findings (Bird 2011, Halkjelsvik 2018) that never test service boundaries. Unblock with a study that measures cost, delivery performance or failures for comparable systems built as microservices and as monoliths, or tracks greenfield microservice projects against monolith-first ones.
Save each document to topics/software-practices/sources/ as text (pdftotext -layout for PDFs), then follow scripts/briefs/sources.md for the new documents against every claim and record with npm run record-sources -- software-practices topics/software-practices/work/sv_*.json.
Then run npm run report and npm run stories -- discover software-practices. If nothing decides it, say so and stop.
Reviewer: The only sources are a review of practitioner reports (Soldani 2018) that lists pains without measuring cost or outcomes, a DORA footnote advising against premature decomposition with no data, and coordination findings (Bird 2011, Halkjelsvik 2018) that never test service boundaries. Unblock with a study that measures cost, delivery performance or failures for comparable systems built as microservices and as monoliths, or tracks greenfield microservice projects against monolith-first ones.
Find studies or public data that would decide this gap in topic heat-pumps, story conflict:heat-pumps-need-more-maintenance-repairs:
Reviewer: the only evidence is seven repair visits on prototype units in the DOE cold-climate challenge (0.2); repair-frequency or maintenance-cost data for many installed heat pumps against gas furnaces, such as a large owner reliability survey or program warranty records, would unblock it
Save each document to topics/heat-pumps/sources/ as text (pdftotext -layout for PDFs), then follow scripts/briefs/sources.md for the new documents against every claim and record with npm run record-sources -- heat-pumps topics/heat-pumps/work/sv_*.json.
Then run npm run report and npm run stories -- discover heat-pumps. If nothing decides it, say so and stop.
Reviewer: the only evidence is seven repair visits on prototype units in the DOE cold-climate challenge (0.2); repair-frequency or maintenance-cost data for many installed heat pumps against gas furnaces, such as a large owner reliability survey or program warranty records, would unblock it
Find studies or public data that would decide this gap in topic software-practices, story surprise:per-change-pull-request-gates-break:
Reviewer: The only counter-evidence is Sadowski et al.'s description of Google (0.4: one highly tuned company and its own tooling, not a pull-request workflow, with no throughput comparison), and DORA 2019 says per-change approval breaks down only for external boards, not teammate review. Unblock with a study measuring lead time or deployment frequency against mandatory per-change peer review across several continuous-delivery organisations, for example DORA data split by review practice.
Save each document to topics/software-practices/sources/ as text (pdftotext -layout for PDFs), then follow scripts/briefs/sources.md for the new documents against every claim and record with npm run record-sources -- software-practices topics/software-practices/work/sv_*.json.
Then run npm run report and npm run stories -- discover software-practices. If nothing decides it, say so and stop.
Reviewer: The only counter-evidence is Sadowski et al.'s description of Google (0.4: one highly tuned company and its own tooling, not a pull-request workflow, with no throughput comparison), and DORA 2019 says per-change approval breaks down only for external boards, not teammate review. Unblock with a study measuring lead time or deployment frequency against mandatory per-change peer review across several continuous-delivery organisations, for example DORA data split by review practice.
Find studies or public data that would decide this gap in topic organic-food, story conflict:organic-labelling-mainly-marketing-tool-charge:
Reviewer: The only evidence is the USDA ERS report's description of the certification, segregation, feed and transition costs behind the premium, which notes that willingness to pay also feeds it; it does not measure how much of the premium is cost and how much is margin. Unblocking it needs a cost study that decomposes organic retail premiums into production, handling and certification costs versus retail and brand margins.
Save each document to topics/organic-food/sources/ as text (pdftotext -layout for PDFs), then follow scripts/briefs/sources.md for the new documents against every claim and record with npm run record-sources -- organic-food topics/organic-food/work/sv_*.json.
Then run npm run report and npm run stories -- discover organic-food. If nothing decides it, say so and stop.
Reviewer: The only evidence is the USDA ERS report's description of the certification, segregation, feed and transition costs behind the premium, which notes that willingness to pay also feeds it; it does not measure how much of the premium is cost and how much is margin. Unblocking it needs a cost study that decomposes organic retail premiums into production, handling and certification costs versus retail and brand margins.
Find studies or public data that would decide this gap in topic organic-food, story conflict:washing-produce-removes-most-pesticide-residue:
Reviewer: The corpus only shows that 57.7% of USDA samples still carried detectable residue after a 15-20 second cold rinse; it never measures how much the rinse removed. Unblocking it needs a study that measures residue before and after household washing methods (rinsing, soaking, baking-soda solutions, peeling) for common pesticides and crops.
Save each document to topics/organic-food/sources/ as text (pdftotext -layout for PDFs), then follow scripts/briefs/sources.md for the new documents against every claim and record with npm run record-sources -- organic-food topics/organic-food/work/sv_*.json.
Then run npm run report and npm run stories -- discover organic-food. If nothing decides it, say so and stop.
Reviewer: The corpus only shows that 57.7% of USDA samples still carried detectable residue after a 15-20 second cold rinse; it never measures how much the rinse removed. Unblocking it needs a study that measures residue before and after household washing methods (rinsing, soaking, baking-soda solutions, peeling) for common pesticides and crops.
Find studies or public data that would decide this gap in topic get-rich-quick, story conflict:demeaning-salaried-jobs-college-sales-tactic:
Reviewer: The corpus has one weak study on business owners' earnings versus salaried workers and nothing on degrees; both halves of the claim need their own evidence.
Save each document to topics/get-rich-quick/sources/ as text (pdftotext -layout for PDFs), then follow scripts/briefs/sources.md for the new documents against every claim and record with npm run record-sources -- get-rich-quick topics/get-rich-quick/work/sv_*.json.
Then run npm run report and npm run stories -- discover get-rich-quick. If nothing decides it, say so and stop.
Reviewer: The corpus has one weak study on business owners' earnings versus salaried workers and nothing on degrees; both halves of the claim need their own evidence.
Find studies or public data that would decide this gap in topic get-rich-quick, story surprise:leverage-lets-small-account-take-large:
Reviewer: The pitch's claim is that a stop-loss sized to the account keeps the dollar risk fixed; the only cited evidence is ESMA's qualitative statement that leverage magnifies losses, fees and margin sensitivity, which does not address stops. It needs a cited verdict on gapping or slippage past stop-loss levels (ESMA's gapping and margin close-out paragraphs in the same document are uncited), or account-level data on realised losses against planned stops.
Save each document to topics/get-rich-quick/sources/ as text (pdftotext -layout for PDFs), then follow scripts/briefs/sources.md for the new documents against every claim and record with npm run record-sources -- get-rich-quick topics/get-rich-quick/work/sv_*.json.
Then run npm run report and npm run stories -- discover get-rich-quick. If nothing decides it, say so and stop.
Reviewer: The pitch's claim is that a stop-loss sized to the account keeps the dollar risk fixed; the only cited evidence is ESMA's qualitative statement that leverage magnifies losses, fees and margin sensitivity, which does not address stops. It needs a cited verdict on gapping or slippage past stop-loss levels (ESMA's gapping and margin close-out paragraphs in the same document are uncited), or account-level data on realised losses against planned stops.
Find studies or public data that would decide this gap in topic get-rich-quick, story contradicted:beginners-experience-can-make-real-money:
Reviewer: The only evidence is an FTC complaint about one firm's 'AI' store buyers; nothing measures beginners' results across AI side hustles.
Save each document to topics/get-rich-quick/sources/ as text (pdftotext -layout for PDFs), then follow scripts/briefs/sources.md for the new documents against every claim and record with npm run record-sources -- get-rich-quick topics/get-rich-quick/work/sv_*.json.
Then run npm run report and npm run stories -- discover get-rich-quick. If nothing decides it, say so and stop.
Reviewer: The only evidence is an FTC complaint about one firm's 'AI' store buyers; nothing measures beginners' results across AI side hustles.
Why do get-rich-quick success stories sound so convincing?
0.46
0.76
0.00
2
Needs research
Next: conflict:advice-people-who-got-rich-quickly
Give an agent this prompt
Find studies or public data that would decide this gap in topic get-rich-quick, story conflict:advice-people-who-got-rich-quickly:
Reviewer: The only study is an agent-based toy model of talent versus luck; it supplies the mechanism, not evidence about advice from people who got rich quickly.
Save each document to topics/get-rich-quick/sources/ as text (pdftotext -layout for PDFs), then follow scripts/briefs/sources.md for the new documents against every claim and record with npm run record-sources -- get-rich-quick topics/get-rich-quick/work/sv_*.json.
Then run npm run report and npm run stories -- discover get-rich-quick. If nothing decides it, say so and stop.
Reviewer: The only study is an agent-based toy model of talent versus luck; it supplies the mechanism, not evidence about advice from people who got rich quickly.
Find studies or public data that would decide this gap in topic get-rich-quick, story contradicted:experienced-day-trader-can-multiply-small:
Reviewer: The corpus has base-rate studies of futures and stock day traders, which cannot rule out an individual momentum trader's month. It needs audited records of momentum traders or a study of account-growth distributions.
Save each document to topics/get-rich-quick/sources/ as text (pdftotext -layout for PDFs), then follow scripts/briefs/sources.md for the new documents against every claim and record with npm run record-sources -- get-rich-quick topics/get-rich-quick/work/sv_*.json.
Then run npm run report and npm run stories -- discover get-rich-quick. If nothing decides it, say so and stop.
Reviewer: The corpus has base-rate studies of futures and stock day traders, which cannot rule out an individual momentum trader's month. It needs audited records of momentum traders or a study of account-growth distributions.
Can you still make money dropshipping or running a faceless channel?
0.44
0.73
0.00
5
Needs research
Next: decision:selling-online
Give an agent this prompt
Find studies or public data that would decide this gap in topic get-rich-quick, story decision:selling-online:
Reviewer: No study in the corpus bears on any claim in this group; the crowding mechanism and the failure rates are AI guesses and source assertions only.
Save each document to topics/get-rich-quick/sources/ as text (pdftotext -layout for PDFs), then follow scripts/briefs/sources.md for the new documents against every claim and record with npm run record-sources -- get-rich-quick topics/get-rich-quick/work/sv_*.json.
Then run npm run report and npm run stories -- discover get-rich-quick. If nothing decides it, say so and stop.
Reviewer: No study in the corpus bears on any claim in this group; the crowding mechanism and the failure rates are AI guesses and source assertions only.
Find studies or public data that would decide this gap in topic get-rich-quick, story gap:technical-patterns-such-fibonacci-retracements-fair:
No study in the corpus checks the claim; the story is the search for evidence.
Save each document to topics/get-rich-quick/sources/ as text (pdftotext -layout for PDFs), then follow scripts/briefs/sources.md for the new documents against every claim and record with npm run record-sources -- get-rich-quick topics/get-rich-quick/work/sv_*.json.
Then run npm run report and npm run stories -- discover get-rich-quick. If nothing decides it, say so and stop.
No study in the corpus checks the claim; the story is the search for evidence.
Do AI code review tools catch real bugs that human reviewers miss?
0.39
0.65
0.00
1
Needs research
Next: gap:code-review-tools-catch-real-bugs
Give an agent this prompt
Find studies or public data that would decide this gap in topic ai-coding, story gap:code-review-tools-catch-real-bugs:
No study in the corpus checks the claim; the story is the search for evidence.
Save each document to topics/ai-coding/sources/ as text (pdftotext -layout for PDFs), then follow scripts/briefs/sources.md for the new documents against every claim and record with npm run record-sources -- ai-coding topics/ai-coding/work/sv_*.json.
Then run npm run report and npm run stories -- discover ai-coding. If nothing decides it, say so and stop.
No study in the corpus checks the claim; the story is the search for evidence.
Find studies or public data that would decide this gap in topic ai-coding, story gap:length-tasks-agents-can-complete-autonomously:
No study in the corpus checks the claim; the story is the search for evidence.
Save each document to topics/ai-coding/sources/ as text (pdftotext -layout for PDFs), then follow scripts/briefs/sources.md for the new documents against every claim and record with npm run record-sources -- ai-coding topics/ai-coding/work/sv_*.json.
Then run npm run report and npm run stories -- discover ai-coding. If nothing decides it, say so and stop.
No study in the corpus checks the claim; the story is the search for evidence.
Are free open-weight models now good enough for everyday coding?
0.37
0.61
0.00
1
Needs research
Next: gap:open-weight-coding-models-now-handle
Give an agent this prompt
Find studies or public data that would decide this gap in topic ai-coding, story gap:open-weight-coding-models-now-handle:
No study in the corpus checks the claim; the story is the search for evidence.
Save each document to topics/ai-coding/sources/ as text (pdftotext -layout for PDFs), then follow scripts/briefs/sources.md for the new documents against every claim and record with npm run record-sources -- ai-coding topics/ai-coding/work/sv_*.json.
Then run npm run report and npm run stories -- discover ai-coding. If nothing decides it, say so and stop.
No study in the corpus checks the claim; the story is the search for evidence.
Are coding agents that run tools themselves really a leap beyond chatting with an AI?
0.36
0.60
0.00
1
Needs research
Next: gap:coding-agents-run-tools-loop-step
Give an agent this prompt
Find studies or public data that would decide this gap in topic ai-coding, story gap:coding-agents-run-tools-loop-step:
No study in the corpus checks the claim; the story is the search for evidence.
Save each document to topics/ai-coding/sources/ as text (pdftotext -layout for PDFs), then follow scripts/briefs/sources.md for the new documents against every claim and record with npm run record-sources -- ai-coding topics/ai-coding/work/sv_*.json.
Then run npm run report and npm run stories -- discover ai-coding. If nothing decides it, say so and stop.
No study in the corpus checks the claim; the story is the search for evidence.
Find studies or public data that would decide this gap in topic software-practices, story gap:pairing-mobbing-puts-best-each-participant:
No study in the corpus checks the claim; the story is the search for evidence.
Save each document to topics/software-practices/sources/ as text (pdftotext -layout for PDFs), then follow scripts/briefs/sources.md for the new documents against every claim and record with npm run record-sources -- software-practices topics/software-practices/work/sv_*.json.
Then run npm run report and npm run stories -- discover software-practices. If nothing decides it, say so and stop.
No study in the corpus checks the claim; the story is the search for evidence.
Find studies or public data that would decide this gap in topic software-practices, story gap:writing-tests-first-forces-code-decoupled:
No study in the corpus checks the claim; the story is the search for evidence.
Save each document to topics/software-practices/sources/ as text (pdftotext -layout for PDFs), then follow scripts/briefs/sources.md for the new documents against every claim and record with npm run record-sources -- software-practices topics/software-practices/work/sv_*.json.
Then run npm run report and npm run stories -- discover software-practices. If nothing decides it, say so and stop.
No study in the corpus checks the claim; the story is the search for evidence.