---
nexus:
  - ora
type: framework
status: active
tags:
  - campaign
  - reproducibility
  - performance
  - framework/instruction
date created: 2026-06-12
date modified: 2026-06-20
related:
  - "[[Paper — Ora Performance]]"
  - "[[Framework — Ora Performance Judging and Scoring]]"
  - "[[Working — Campaign Run Plan]]"
  - "[[Reference — Trigger Prompt Corpus]]"
---

# Framework — Campaign Self-Verification

## Display Name

Ora Performance Capture Reproduction

## Display Description

Runs the Ora Performance capture campaign from a self-contained prompt library using API calls only. Produces accepted captures, configuration snapshots, timing, cost, and audit records for the six configurations used in the paper.

## Purpose

This framework lets an independent operator reproduce the capture side of the Ora Performance comparison without using the author's subscription account, private vault, or local campaign notes. The output is not the final score table; it is the capture bundle that the separate judging framework scores.

The campaign covers **196 entries**: 60 analysis modes, 21 visual outputs, and 115 lenses. All 60 modes are scored, including `passion-exploration` and `structured-output` (both captured on fewer than all six lanes in the reference run). Two techniques share a public ID across kinds — `causal-dag` (analytical mode and visual tool) and `fishbone-diagram` (interpretive lens and visual tool) — and the reference run scored one entry per shared ID: the analytical-mode prompt for `causal-dag` and the visual-tool prompt for `fishbone-diagram`. So the visual `causal-dag` entry and the lens `fishbone-diagram` entry are intentionally absent here.

## Setup Questions

### Ora execution environment

Required. Provide access to an Ora install or API-compatible implementation of the Ora harness. It must be able to run the Gear 4 pipeline, the model picker, visual generation for visual entries, and bare one-call controls.

### API credentials

Required. Provide metered API access for every model used in the campaign. Do not use subscription accounts, unlimited-plan accounts, free hidden quota, or manually operated web chats. If a provider supports both subscription and API access, use API access.

### Run scale

Required. Choose one: smoke test, selected subset, or full 196-entry campaign. A full run can take many days and should be checkpointed after every accepted capture.

### Output location

Required. Choose a directory where the operator can persist configuration snapshots, captures, visual artifacts, token/cost records, wall-clock timings, acceptance decisions, and errors.

### Local 9B availability

Required. State whether the 9B lane will be served from a local on-device model or through a metered API endpoint for the same model. Either is acceptable if the model identity is recorded and the run uses the same model in all small-local roles.

### Model catalog date

Required. Record the date and source of the available model catalog. Model catalogs and prices move; the result is reproducible only if the selected model IDs and rates are frozen in the output bundle.

## Intake Protocol

1. Restate the run scope: number of entries, configurations, expected captures, output location, and whether this is a smoke test, subset, or full run.
2. Confirm the run is API-only. If any lane would use a subscription account or web-chat account, stop and require an API route instead.
3. Confirm the scored library contains 196 entries: all 60 analysis modes (including `passion-exploration` and `structured-output`), 21 visual outputs, and 115 lenses, with each shared-ID collision resolved to one entry (`causal-dag` as a mode, `fishbone-diagram` as a visual).
4. Confirm the environment has no private retrieval enabled. The campaign may use public web retrieval when the harness normally would, but it must not use the operator's private vault, private conversation memory, or unpublished notes.
5. Run the model picker before capture, then freeze the selected model IDs, provider IDs, and prices. Do not allow later catalog changes to rewrite the snapshot.
6. Explain the likely cost/time envelope before starting. A full run is expensive and slow because it repeats every prompt across six configurations.

## Execution

### Phase 1 — Build the Configuration Snapshot

Create six named configurations and record the exact model IDs, providers, rates, and role assignments used by each.

1. **Flagship Model** — one bare API call to the flagship model used by the Premium configuration's primary large-model role. No harness, no tools, no review, no reruns, no private memory.
2. **Small Model, Raw** — one bare API call to the selected 9B model. No harness, no tools, no review, no reruns, no private memory.
3. **Small Local Only** — the selected 9B model in every Ora harness role: large-model roles, fast roles, small role, utility, consolidation, and verification. If the 9B is served by API rather than locally, record that fact; do not describe it as on-device.
4. **Optimum** — run the Ora model picker under the cost-aware optimum policy. Record every role: Big 1, Big 2, Fast 1, Fast 2, Small, Utility, Consolidate, and Verify.
5. **Optimum + Premium** — duplicate the frozen Optimum configuration, then replace only the final consolidation role with the Premium configuration's primary flagship model. Leave the research, review, verification, utility, and other pipeline roles unchanged.
6. **Premium** — run the Ora model picker under the maximum-capability policy using API-accessible premium/frontier models throughout. Record the large, fast, small, utility, consolidation, and verification roles.

If the picker cannot produce a required configuration, stop and report the missing model class. Do not silently substitute a cheaper or more available model after the snapshot is frozen.

### Phase 2 — Prepare Each Campaign Entry

For each entry in the prompt library:

1. Use the qualified key as the canonical identifier: `mode:<id>`, `visual:<id>`, or `lens:<id>`.
2. Use the listed prime prompt exactly. Do not rewrite, shorten, current-date-adjust, or select another prompt.
3. Pin the intended operation. Modes pin their intended mode; visuals pin the routed mode and expected visual type; lenses pin the host mode and foregrounded lens.
4. Record whether the entry is a mode, visual output, or lens.
5. Resolve the two shared-ID collisions to a single entry each, as the reference run did: score `causal-dag` as a mode (not also as a visual output) and `fishbone-diagram` as a visual output (not also as a lens). The library below already reflects this; do not re-add the dropped entries.

### Phase 3 — Capture Answers

For each campaign entry, run all six configurations against the same prime prompt.

For raw configurations, make exactly one API call to the configured model and capture the returned answer, usage, latency, and cost.

For harnessed configurations, run the Ora harness normally for the pinned operation. Capture the final answer, any visual artifact, model calls by role, token usage, cost, wall-clock time, review/rerun decisions, and any operational warnings.

Checkpoint after every accepted capture. A full campaign is allowed to span many sessions as long as the manifest prevents duplicate accepted captures and makes resumed work explicit.

### Phase 4 — Acceptance Gate

Before a capture counts, verify all of the following:

1. The exact frozen model IDs served the roles they were assigned to serve.
2. No fallback model, default model, or manual web-chat model answered in place of a configured API model.
3. The prompt was the library prime prompt, unchanged.
4. The operation stayed pinned to the intended mode, visual route, or lens host.
5. Visual entries produced and preserved the visual artifact when the harness normally emits one. Raw one-call controls may be text-only, but that absence must be recorded.
6. The answer did not draw on the operator's private vault, private conversation memory, unpublished notes, or hidden manual context.
7. Costs and timings were recorded from the actual run, not estimated after the fact.

If a capture fails the gate, rerun it when the problem is transient. If the problem is structural, mark the capture invalid and report why. Invalid captures are not scored as model failures.

### Phase 5 — Persist the Capture Bundle

For every accepted capture, persist:

- qualified entry key
- family: mode, visual, or lens
- intended mode, visual route, or lens host
- exact prime prompt
- configuration name
- frozen model role map
- final answer
- visual artifact files or explicit text-only note
- token usage by call when available
- API cost by call and total cost
- wall-clock duration
- acceptance-gate result
- error and rerun history

Also persist a campaign-level configuration snapshot, cost summary, timing summary, completion report, and manifest.

### Phase 6 — Produce the Reproducibility Report

At the end of the selected run scope, produce a short report with:

1. The model catalog date and every frozen model ID.
2. The number of accepted captures by configuration.
3. The number of invalid or failed captures, separated from model-quality failures.
4. Total and per-configuration API cost.
5. Average wall-clock time per configuration.
6. Any deviations from this framework.
7. A clear statement that scoring requires the separate judging framework.

## Output Contract

This framework produces a capture bundle, not a judgment. A valid bundle contains accepted answers for the selected entries, model/configuration snapshots, prompt provenance, timing/cost records, visual artifacts where applicable, and acceptance-gate decisions. The full campaign grid is 196 entries x 6 configurations = 1,176 possible captures; the reference run holds 1,173 accepted (the modes `passion-exploration` and `structured-output` captured on fewer than all six lanes).

## Failure Modes

### Subscription Contamination

A run uses a subscription account, web chat, or hidden quota. Reject the affected captures and rerun through metered APIs.

### Silent Model Substitution

A provider serves a fallback or default model instead of the frozen model. Reject the capture; do not count it as model behavior.

### Collision-ID double count

Both entries of a shared-ID collision enter the scored set — e.g. `causal-dag` scored as both a mode and a visual output, or `fishbone-diagram` as both a lens and a visual. Score one entry per shared ID (the mode for `causal-dag`, the visual for `fishbone-diagram`) and keep the denominator at 196.

### Private Context Leakage

The harness consults private vault material, conversation memory, or unpublished notes. Reject the capture and rerun with private retrieval disabled.

### Prompt Drift

A prompt is rewritten, shortened, or replaced with a non-prime example. Reject the capture and rerun with the exact library prompt.

### Configuration Drift

The model picker is rerun mid-campaign and changes the configuration. Keep the original frozen snapshot for the campaign; only start a new campaign if the snapshot changes.

### Cost Retrofitting

Costs are estimated after the run from current prices instead of captured at runtime. Mark the cost data non-reproducible and rerun if cost comparison matters.

## Appendix — Minimal Manifest Schema

Each capture row should contain these fields or their exact equivalents:

```yaml
entry_key: mode:argument-audit
family: mode
intended_operation: argument-audit
prime_prompt: "..."
configuration: optimum
model_snapshot_id: 2026-06-20T00:00:00Z
models_used:
  analyst: provider/model-id
  reviewer: provider/model-id
  consolidator: provider/model-id
answer_path: captures/mode-argument-audit/optimum/answer.md
visual_paths: []
api_cost_usd: 0.00
wall_clock_seconds: 0
accepted: true
acceptance_notes: "exact configured models served; no fallback"
```

## Prompt Library — 196 Entries

This library is part of the framework. Do not require the operator to download `Reference — Trigger Prompt Corpus.md` separately. Use prompt 1 only for every entry. All 60 analysis modes are included. Each shared-ID collision is resolved to one entry: `causal-dag` appears as a mode (not also as a visual output) and `fishbone-diagram` as a visual output (not also as a lens), matching the reference run.

### Modes — 60 entries

#### mode:argument-audit
- **Intended operation:** `argument-audit`
- **Prime prompt:** Argument audit on this: 'AI alignment is solved because RLHF works. We've trained models to be helpful, harmless, and honest, so the problem reduces to scaling that training. The remaining work is engineering, not research.' Coherence and frame check together.

#### mode:coherence-audit
- **Intended operation:** `coherence-audit`
- **Prime prompt:** I keep hearing 'tariffs don't cause inflation because they're a one-time price-level shift, not ongoing.' Does this argument hold up?

#### mode:frame-audit
- **Intended operation:** `frame-audit`
- **Prime prompt:** Frame audit on this Atlantic Council dispatch about the Anthropic-Pentagon standoff: 'Treating public skepticism as noise to be managed rather than a signal to be heeded risks causing rapid political polarization on AI. The standoff exposed the fraying social contract among leading AI companies, the federal government, and the American public over responsible AI use. The administration's maximalist position that contracts with AI companies should provide flexibility for the government to employ AI for all lawful uses runs counter to US public opinion. For the administration's and the tech sector's AI ambitions to translate into economic growth and national security gains, it will take a concerted effort to rebuild the social contract with the public on AI.' What is selected in and selected out? What frame is this using?

#### mode:propaganda-audit
- **Intended operation:** `propaganda-audit`
- **Prime prompt:** Propaganda audit, Stanley test, supporting vs undermining propaganda on this Workers Party of Britain manifesto excerpt: 'The greatest block to working class aspirations is not the Conservative Party but the Labour party itself. Labour is nothing more than a wolf in sheep's clothing, an integrated part of the imperialist State machinery. We do not hide our belief.' Flawed-ideology premise check, manufactured-doubt diagnostic.

#### mode:boundary-critique
- **Intended operation:** `boundary-critique`
- **Prime prompt:** Boundary critique of the new AI Safety Institute charter — whose voice is missing, who is excluded from the definition of 'public interest'?

#### mode:cui-bono
- **Intended operation:** `cui-bono`
- **Prime prompt:** Who benefits from the EU's new AI Act timeline being pushed to 2027? Trace the interests — feels like someone's pushing this for a reason.

#### mode:decision-clarity
- **Intended operation:** `decision-clarity`
- **Prime prompt:** Decision clarity document for the board: should our school district adopt year-round schooling? Brief the decision-maker, not exploratory analysis.

#### mode:wicked-problems
- **Intended operation:** `wicked-problems`
- **Prime prompt:** Homelessness in our mid-sized city is a wicked problem. Every solution we try makes it worse somewhere else: shelters concentrate people, housing-first runs into permitting, encampment sweeps push folks further out. Stakeholders disagree about what the problem even is. Everything is connected. Where do we start?

#### mode:constraint-mapping
- **Intended operation:** `constraint-mapping`
- **Prime prompt:** Compare alternatives for our cloud migration — AWS-only, multi-cloud, hybrid. Map the tradeoffs across cost, lock-in, and team skill.

#### mode:decision-architecture
- **Intended operation:** `decision-architecture`
- **Prime prompt:** Full decision analysis on a real decision we're facing this quarter: 'Accept the $40M Series B at flat valuation now (preferred terms, lead investor wants 8 weeks to close), OR bridge with a $5M convertible note from existing investors and target an $80M Series B in six months once our next two enterprise deals close.' Runway 14 months either way. Structured decision document — stakeholders, options, criteria, integrated. Decision architecture.

#### mode:decision-under-uncertainty
- **Intended operation:** `decision-under-uncertainty`
- **Prime prompt:** Decision under uncertainty: should we exercise the office-lease purchase option now or wait six months? Building $4.2M, mortgage 7.1%, lease option expires in eight months and refunds 50% if not exercised. Downside of each option matters — capital tied up vs. losing the option entirely. Decision tree with expected value across rate-cut and rate-flat outcomes, value of information on waiting one more Fed meeting. Should we act now or wait?

#### mode:multi-criteria-decision
- **Intended operation:** `multi-criteria-decision`
- **Prime prompt:** Multi-criteria decision on choosing a graduate program — rank options across reputation, cost, location, fit. AHP weighting.

#### mode:causal-dag
- **Intended operation:** `causal-dag`
- **Prime prompt:** I want to know what would happen if we intervened — raising minimum wage by 25% in our state. Causal DAG that separates confounders from the causal path.

#### mode:process-tracing
- **Intended operation:** `process-tracing`
- **Prime prompt:** What really happened with the 2026 Indian Ocean cable cut? Process tracing — smoking gun, hoop test, which causal story does the evidence support?

#### mode:root-cause-analysis
- **Intended operation:** `root-cause-analysis`
- **Prime prompt:** What are the root causes of our team's missed sprint deadlines the last three quarters? Draw a fishbone.

#### mode:systems-dynamics-causal
- **Intended operation:** `systems-dynamics-causal`
- **Prime prompt:** Why does fixing X make things worse: every time we hire more customer-support agents, response time improves briefly then degrades worse than before. Counterintuitive results, a feedback loop somewhere we can't see. Diagnose the recurring behaviour.

#### mode:bayesian-hypothesis-network
- **Intended operation:** `bayesian-hypothesis-network`
- **Prime prompt:** Build me a Bayesian network on what's driving the 2026 housing slowdown — rates, demographics, remote-work reversal. Set up priors and update with evidence properly.

#### mode:competing-hypotheses
- **Intended operation:** `competing-hypotheses`
- **Prime prompt:** I have three competing hypotheses for why our app engagement dropped 30% in March: algorithm change, content fatigue, seasonal effect. Make me an ACH matrix.

#### mode:differential-diagnosis
- **Intended operation:** `differential-diagnosis`
- **Prime prompt:** My laptop fan has been running constantly for two days. Battery drains in three hours instead of eight. CPU shows 40% at idle. Candidate explanations: runaway background process, failing battery, malware, thermal-paste degradation, dust buildup. Differential diagnosis — which is most likely?

#### mode:consequences-and-sequel
- **Intended operation:** `consequences-and-sequel`
- **Prime prompt:** If we ship this — OpenAI's Workspace Agents (announced April 22 2026): persistent autonomous agents inside enterprise suites with admin-controlled permissions, credit-based pricing, scheduled long-running tasks — rolled out to the full enterprise customer base over two quarters, what does this lead to? What are the downstream effects, the cascade forward over a longer horizon?

#### mode:pre-mortem-action
- **Intended operation:** `pre-mortem-action`
- **Prime prompt:** Pre-mortem this plan: we're launching a new pricing page next Tuesday — three tiers, annual discount, redesigned hero. What kills it?

#### mode:probabilistic-forecasting
- **Intended operation:** `probabilistic-forecasting`
- **Prime prompt:** What's the probability that Apple ships a folding iPhone in 2027? Calibrated estimate with base rate for major form-factor pivots.

#### mode:scenario-planning
- **Intended operation:** `scenario-planning`
- **Prime prompt:** Scenario planning for our retirement portfolio over the next 15 years — 2x2 scenario matrix across rates and inflation.

#### mode:wicked-future
- **Intended operation:** `wicked-future`
- **Prime prompt:** Wicked future on AI labor displacement over 2026-2031: long-horizon scenarios with probabilities, plus pre-mortem failure pathways.

#### mode:pre-mortem-fragility
- **Intended operation:** `pre-mortem-fragility`
- **Prime prompt:** Pre-mortem this design: our new auth service architecture — three microservices, shared JWT, Redis session store. Where will this break? Single points of failure.

#### mode:fragility-antifragility-audit
- **Intended operation:** `fragility-antifragility-audit`
- **Prime prompt:** Fragility audit on our supply chain: where are the tail-risk exposures, convex or concave, asymmetric payoffs? Taleb-style.

#### mode:stakeholder-mapping
- **Intended operation:** `stakeholder-mapping`
- **Prime prompt:** Stakeholder map for our hospital merger using Mitchell Agle Wood salience. RACI for the integration committee. Identifying who's involved, who has standing, and what each party's stake is — board members, union nurses, regulators, insurers. Who else needs to be at the table?

#### mode:frame-comparison
- **Intended operation:** `frame-comparison`
- **Prime prompt:** Frame comparison on grid decarbonization. The climate-policy frame: 'phase out gas peakers by 2030, build storage and transmission, accept short-term reliability risk for long-term emissions.' The energy-realist frame: 'keep gas as firm capacity, build renewables on top, retire only when storage is proven at grid scale.' Compare both worldviews on their own terms.

#### mode:paradigm-suspension
- **Intended operation:** `paradigm-suspension`
- **Prime prompt:** Suspend the paradigm — what if the consensus that 10,000 daily steps is the health threshold is wrong? Heterodox exploration.

#### mode:worldview-cartography
- **Intended operation:** `worldview-cartography`
- **Prime prompt:** Worldview cartography of the AI safety debate. Four paradigms: existential-risk doomers (Yudkowsky lineage), accelerationists / e/acc, alignment-pragmatists (Anthropic-style), and the AI-skeptic / pattern-matching camp. Multi-paradigm map — where do they cohere, where do they irreducibly conflict?

#### mode:conceptual-engineering
- **Intended operation:** `conceptual-engineering`
- **Prime prompt:** Conceptual engineering on 'consent' as used in current data-privacy law — the inherited concept isn't doing the work it should. What should it mean?

#### mode:deep-clarification
- **Intended operation:** `deep-clarification`
- **Prime prompt:** Explain the mechanics of how transformer attention actually works — I want depth, the math internals, not orientation.

#### mode:relationship-mapping
- **Intended operation:** `relationship-mapping`
- **Prime prompt:** Relationship map of how the new federal AI regulation interacts with state-level laws — dependency graph, what affects what.

#### mode:spatial-reasoning
- **Intended operation:** `spatial-reasoning`
- **Prime prompt:** Annotate my concept map of organizational change. Nodes: leadership-buy-in, middle-management-resistance, employee-engagement, training-budget, communication-cadence, success-metrics, culture-shift. Edges drawn: leadership→training-budget, training-budget→employee-engagement, communication-cadence→engagement, engagement→culture-shift. What node am I missing? Is there a feedback loop I haven't drawn?

#### mode:dialectical-analysis
- **Intended operation:** `dialectical-analysis`
- **Prime prompt:** Dialectical analysis: tech innovation vs. precautionary regulation. Both have real merit, compromise feels like cop-out. Drive through the contradiction.

#### mode:synthesis
- **Intended operation:** `synthesis`
- **Prime prompt:** Synthesize two bodies of knowledge developed separately: Buddhist Vipassana meditation (anatta, dependent origination, mindfulness of feeling tone) and Western cognitive-behavioral therapy (cognitive distortions, behavioral activation, exposure). What's the structural parallel? Map the intersection.

#### mode:interest-mapping
- **Intended operation:** `interest-mapping`
- **Prime prompt:** Interest mapping before I enter a salary negotiation with my new manager. What does each side really want, separate from positions?

#### mode:principled-negotiation
- **Intended operation:** `principled-negotiation`
- **Prime prompt:** Principled negotiation prep for buying a small business — BATNA, options for mutual gain, objective criteria. Full Fisher-Ury treatment.

#### mode:third-side
- **Intended operation:** `third-side`
- **Prime prompt:** Third-side reading of this dispute: dock workers' union and port management have been deadlocked for six weeks. Mayor's office, two community coalitions, and the state DOT are circling. I'm advising the mayor on her mediation role. What third-side roles are available — Ury's ten roles, the community's role in containing this?

#### mode:domain-induction
- **Intended operation:** `domain-induction`
- **Prime prompt:** Induct me into this domain: residential structural engineering for kitchen remodels. I'm a homeowner who needs structured introduction to a domain — what's here, what's connected to what, and the ordered learning pathway with prerequisites. Comprehensive domain introduction.

#### mode:quick-orientation
- **Intended operation:** `quick-orientation`
- **Prime prompt:** Give me the quick lay of the land on graph databases — I have ten minutes.

#### mode:terrain-mapping
- **Intended operation:** `terrain-mapping`
- **Prime prompt:** I'm unfamiliar with monetary policy. Walk me through the big picture — concept map of how the Fed actually shapes the economy.

#### mode:balanced-critique
- **Intended operation:** `balanced-critique`
- **Prime prompt:** Balanced critique of the UK 4-Day Week Pilot Programme final results, which reported: '35% average increase in revenue,' '57% decrease in attrition rate,' '71% decrease in employee burnout,' '92% of businesses plan to continue,' '55% reported an increase in work ability.' 61 companies, 32 hours/week, six months, no wage reduction. What holds up and what doesn't — strengths and weaknesses, no advocacy.

#### mode:benefits-analysis
- **Intended operation:** `benefits-analysis`
- **Prime prompt:** What are the benefits and risks of this proposal I'm weighing: 'Migrate our 80-person engineering team from Slack to Discord over Q3 2026. Reasons cited: 40% Slack-bill cost saving, native voice rooms for pair-programming, better threading, async-by-default culture fit. Concerns raised internally: enterprise compliance, search/audit, integration loss with Linear and Notion.' Full picture on this one option — pros and cons, evaluate this proposal, no comparing alternatives.

#### mode:red-team-advocate
- **Intended operation:** `red-team-advocate`
- **Prime prompt:** Argue against OpenAI's $1.5B DeployCo joint venture (announced April 2026 with TPG-led PE consortium). The strategic rationale: 'OpenAI plans to take a percentage of the value created or costs saved by its agentic workflows,' deploying 'forward deployed engineers' inside client orgs to 'rewire' business processes, with a 17.5% guaranteed annual return to PE backers and access to 1,200+ portfolio companies. Stated goal: vertical integration to close the enterprise 'ROI gap' against Anthropic and Google. Make the case against this strategy — every angle including weak ones, no severity triage, comprehensive critique.

#### mode:red-team-assessment
- **Intended operation:** `red-team-assessment`
- **Prime prompt:** Stress-test this go-to-market plan for our Series-A B2B SaaS launch before I commit. Plan: 'Three coastal cities first (NYC, SF, LA). $200/mo single-tier subscription, no free trial, 90-day money-back guarantee. Paid acquisition Meta + TikTok at $80 CAC target. Outbound SDRs in month 3. ICP: 50-500 person ops teams. Public launch month 6 after closed beta of 50 customers at founder rates.' Where is this weak? What am I missing? Find the holes — what would a hostile reviewer say?

#### mode:steelman-construction
- **Intended operation:** `steelman-construction`
- **Prime prompt:** Steelman this published position from Senator Sanders' op-ed — best case for, strongest version: 'In the richest country on earth, if you work 40 hours a week you do not live in poverty. Putting money into the hands of people who will spend it on basic needs is a strong economic stimulant. When over 60% of American workers are now living paycheck to paycheck, we can no longer tolerate a federal minimum wage of $7.25 an hour. Poll after poll shows overwhelming support for raising the minimum wage to a living wage. There are too many Americans trying to survive on $9, $10 or $12 an hour. It cannot be done.' No caveat-padding.

#### mode:mechanism-understanding
- **Intended operation:** `mechanism-understanding`
- **Prime prompt:** How does CRISPR actually work under the hood? Want the structural explanation — how the parts produce the editing behavior, not just inputs and outputs.

#### mode:market-dynamics
- **Intended operation:** `market-dynamics`
- **Prime prompt:** Walk me through the supply and demand in the US used-EV market after the 2026 expiration of the federal clean-vehicle tax credit. What happens to prices and volumes in the short run versus the long run once the subsidy is gone, and where does the market settle? Elasticities on both sides, and the new equilibrium.

#### mode:process-mapping
- **Intended operation:** `process-mapping`
- **Prime prompt:** Process map for our customer onboarding workflow — find the bottlenecks and dependencies. Current state, as-is.

#### mode:systems-dynamics-structural
- **Intended operation:** `systems-dynamics-structural`
- **Prime prompt:** Draw the feedback structure (structural) for our org's promotion cycle. Show me the stocks (eligible candidates, open seats, retained ICs), the flows (hiring, promotion, attrition), and a structural diagram with feedback dynamics. I want a structural picture before deciding what to change.

#### mode:mechanism-design
- **Intended operation:** `mechanism-design`
- **Prime prompt:** We're designing a procurement auction for municipal construction contracts and keep getting burned by the winner's curse and lowball bids from firms that then cut corners. Analyze the adverse-selection and moral-hazard problems here, then design the auction rules and contract incentives so honest, capable bidders self-select and win. Mechanism design — screening and incentive-compatibility.

#### mode:strategic-interaction
- **Intended operation:** `strategic-interaction`
- **Prime prompt:** Game theory: what's China's best move if the US imposes export controls on advanced chips in Q3? Payoff matrix, deterrence, signaling.

#### mode:compositional-dynamics
- **Intended operation:** `compositional-dynamics`
- **Prime prompt:** Compositional dynamics reading of the new MoMA exhibition poster — figure-ground, gestalt grouping, where the eye goes and why.

#### mode:information-density
- **Intended operation:** `information-density`
- **Prime prompt:** Critique this chart with a Tufte data-ink-ratio audit: a quarterly revenue dashboard with 3D pie chart, drop shadows, rainbow gradient bars for each region, gridlines every $1k, a sparkline-sized YoY indicator with no axis labels, and a legend taking 30% of the canvas. Chartjunk audit, Bertin visual-variables read, prescriptive recommendations.

#### mode:ma-reading
- **Intended operation:** `ma-reading`
- **Prime prompt:** Ma reading of this Japanese tea-house garden photograph — what is the empty space doing? Yūgen or wabi-sabi at work?

#### mode:place-reading-genius-loci
- **Intended operation:** `place-reading-genius-loci`
- **Prime prompt:** Place reading of the new High Line extension — genius loci, prospect-refuge, Christopher Alexander pattern language. What does the space afford?

#### mode:project-mode
- **Intended operation:** `project-mode`
- **Prime prompt:** Build me a one-page React app that displays my GitHub stars sorted by language.

#### mode:passion-exploration
- **Intended operation:** `passion-exploration`
- **Prime prompt:** I've been wondering about why we find certain landscapes restorative. Let me explore — no deliverable in mind.

#### mode:structured-output
- **Intended operation:** `structured-output`
- **Prime prompt:** Format this trip itinerary as a one-pager with time, location, and notes columns. Faithful rendering of what I've drafted.

### Visual Outputs — 21 entries

#### visual:ach-matrix
- **Routes to mode:** `competing-hypotheses`
- **Prime prompt:** Make me an ACH matrix on three competing hypotheses for the 2026 GDP slowdown: monetary tightening, demand exhaustion, supply normalization. Evidence rows: bond spreads, unemployment, inventory, consumer sentiment. What rules out each theory?

#### visual:bow-tie-diagram
- **Routes to mode:** `fragility-antifragility-audit`
- **Prime prompt:** Bow-tie diagram for our largest cybersecurity risk — ransomware. Threats, preventive controls, top event, consequences, mitigation controls. Fragility audit on the structure, single points of failure.

#### visual:c4-architecture
- **Routes to mode:** `spatial-reasoning`
- **Prime prompt:** Annotate this architecture: I have a C4 context diagram for our payments service — actors (customer, support agent, fraud analyst), our payments system, external systems (Stripe, Plaid, accounting). What node am I missing? Is there a connection I haven't drawn?

#### visual:causal-loop-diagram
- **Routes to mode:** `systems-dynamics-causal`
- **Prime prompt:** Why does fixing X make things worse — every feature release improves retention briefly then worsens it. Counterintuitive results, feedback loop somewhere. Draw the CLD with reinforcing and balancing loops, diagnose the recurring behaviour.

#### visual:comparison-chart
- **Routes to mode:** `information-density`
- **Prime prompt:** Critique this dashboard chart and recommend the right encoding: I have 2024 vs 2025 vs 2026 Q1 revenue across our four product lines. Data-ink ratio, chartjunk, Cleveland-McGill elementary perceptual tasks — what visual encoding is right for this comparison? Bertin visual variables analysis.

#### visual:concept-map
- **Routes to mode:** `terrain-mapping`
- **Prime prompt:** Map this domain for me with a concept map — attachment theory: secure, anxious, avoidant, disorganized. Walk me through the propositions connecting them. The big picture, introduce me to it.

#### visual:decision-tree
- **Routes to mode:** `decision-under-uncertainty`
- **Prime prompt:** Build me a decision tree for the lease-vs-buy decision: lease (certain $50k/yr), buy with 30% down (uncertain appreciation), wait six months (uncertain market). Expected value, value of information.

#### visual:distribution-plot
- **Routes to mode:** `information-density`
- **Prime prompt:** I'm designing this chart and need to choose the right encoding. The data: API response times from last week, ~5M requests, p50 of 80ms but a long tail to 4 seconds. The decision the chart needs to support: where to set our SLO. Critique this chart — box plot, violin, histogram, or strip? Cleveland-McGill graphical perception, data-ink ratio.

#### visual:fishbone-diagram
- **Routes to mode:** `root-cause-analysis`
- **Prime prompt:** Why does this keep happening — our deployment failures recur every release. Draw a fishbone / Ishikawa diagram with categories: people, process, technology, data, environment. What are the root causes?

#### visual:flowchart
- **Routes to mode:** `process-mapping`
- **Prime prompt:** Process map with swimlanes — flow chart for our incident-response workflow: customer support, on-call engineer, SRE, post-mortem owner. Current state, as-is, find the bottleneck.

#### visual:heatmap
- **Routes to mode:** `information-density`
- **Prime prompt:** I'm designing this dashboard chart and need prescriptive recommendations: a heatmap of feature usage by user cohort across 12 weeks, with cohorts on rows, weeks on columns, usage-intensity as fill. Message: power-user cohorts retain feature use, casual cohorts decay by week 4. Cleveland-McGill perceptual-task fit, Bertin visual variables, data-ink audit. Is heatmap the right encoding here?

#### visual:ibis-argument
- **Routes to mode:** `dialectical-analysis`
- **Prime prompt:** Wicked problem — our team's Kubernetes-adoption debate keeps shifting as we try to define it. Stakeholders disagree about what the problem even is. Build me an IBIS issue map with positions and arguments, dialectical analysis of the contradictions.

#### visual:influence-diagram
- **Routes to mode:** `decision-architecture`
- **Prime prompt:** Full decision analysis with a Howard-Matheson influence diagram for our facility-location decision: decision nodes, chance nodes, value node. Decision architecture.

#### visual:pro-con-tree
- **Routes to mode:** `benefits-analysis`
- **Prime prompt:** Pro-con tree for: should we open-source our core library? PMI — plus, minus, interesting. Full picture on this one option, with sub-pros and sub-cons.

#### visual:quadrant-matrix
- **Routes to mode:** `constraint-mapping`
- **Prime prompt:** Show me a 2x2 matrix for these tasks by urgency and importance — Eisenhower-style. Tasks: ship v3 release, fix auth bug, plan Q3 roadmap, review three PRs, write board update. Map the tradeoffs.

#### visual:scatter-plot
- **Routes to mode:** `information-density`
- **Prime prompt:** I'm evaluating this info-graphic and need prescriptive recommendations: a scatter plot of customer lifetime value vs acquisition cost across 200 enterprise accounts, with sales-region encoded as color. Message: high-CAC accounts in the Northeast are underperforming. Tufte data-ink critique, Bertin visual-variables analysis, chartjunk audit.

#### visual:sequence-diagram
- **Routes to mode:** `process-mapping`
- **Prime prompt:** Process map as a Mermaid sequence diagram for the OAuth 2.0 authorization code flow with PKCE: client, browser, auth server, resource server. Step by step how does this work?

#### visual:state-diagram
- **Routes to mode:** `process-mapping`
- **Prime prompt:** Process map as a state diagram for our subscription lifecycle: trial, active, past-due, canceled, won-back. Step by step how does this work — current state.

#### visual:stock-and-flow
- **Routes to mode:** `systems-dynamics-structural`
- **Prime prompt:** Stock and flow diagram for our cash position: revenue inflow, payroll outflow, fundraising inflow, opex outflow. Map the system's loops and flows. Senge archetype if visible.

#### visual:time-series
- **Routes to mode:** `information-density`
- **Prime prompt:** I'm designing this chart and need to choose the right encoding: monthly active users from Jan 2024 through Apr 2026, with a 7-day moving average overlaid. The message is that growth has decelerated since Q3 2025. Critique this chart — Tufte data-ink ratio, sparkline option, small multiples consideration. What visual encoding fits this elementary perceptual task?

#### visual:tornado-chart
- **Routes to mode:** `decision-under-uncertainty`
- **Prime prompt:** Tornado diagram on our DCF valuation — sensitivity on growth rate, discount rate, terminal multiple, churn. Which input variables drive expected value most? Value of information.

### Lenses — 115 entries

#### lens:affect-heuristic
- **Foregrounded lens:** `affect-heuristic`
- **Host mode:** `propaganda-audit`
- **Prime prompt:** Analyze the affect heuristic in how this ad gets people to judge the product's safety by the warm feeling it evokes rather than any data.

#### lens:anchoring
- **Foregrounded lens:** `anchoring`
- **Host mode:** `propaganda-audit`
- **Prime prompt:** Do a propaganda-style audit of the manipulation in this ad: a $999 'original price' slashed to $399. Show how the anchor distorts what buyers think the product is really worth.

#### lens:availability-heuristic
- **Foregrounded lens:** `availability-heuristic`
- **Host mode:** `propaganda-audit`
- **Prime prompt:** After a plane crash dominates the news, people overestimate flying risk and drive instead. Analyze the availability heuristic — judging probability by how easily vivid examples come to mind.

#### lens:choice-architecture
- **Foregrounded lens:** `choice-architecture`
- **Host mode:** `frame-audit`
- **Prime prompt:** Analyze the choice architecture of our checkout flow — how the arrangement of options and the defaults steer what users end up picking.

#### lens:commitment-consistency
- **Foregrounded lens:** `commitment-consistency`
- **Host mode:** `propaganda-audit`
- **Prime prompt:** Propaganda audit on this sales script: 'You'd agree protecting your family matters, right? ... And you'd want the best protection available, wouldn't you? ... So it only makes sense to go with our premium plan today.' Audit the commitment-and-consistency tactic — the small initial yeses engineered so the customer stays consistent with a much bigger commitment. Engineering-of-consent diagnostic.

#### lens:entman-framing-functions
- **Foregrounded lens:** `entman-framing-functions`
- **Host mode:** `frame-audit`
- **Prime prompt:** Frame audit on this editorial line about the protest: 'Violent agitators shut down the city last night, vandalizing storefronts and clashing with police trying to keep order.' What frame is this using? Lakoff/Entman treatment — what's foregrounded, what's backgrounded, what tacit assumptions in this frame; how it defines the problem, diagnoses the cause, renders the moral judgment, and prescribes the remedy.

#### lens:goffman-frame-analysis
- **Foregrounded lens:** `goffman-frame-analysis`
- **Host mode:** `frame-audit`
- **Prime prompt:** Frame audit on this satirical ad using Goffman's frame analysis: what primary framework is it asking the audience to apply, and where is it 'keyed' into a joke not meant literally? Identify the frame, the keying, and where the reading breaks down between audiences who take it straight and those who get the satire.

#### lens:scarcity
- **Foregrounded lens:** `scarcity`
- **Host mode:** `propaganda-audit`
- **Prime prompt:** Propaganda audit on this checkout banner: 'Only 3 left in stock — 14 people are viewing this right now. Order within 9:58 to get it by Friday.' Audit the scarcity tactic — how the manufactured urgency and false-limit cues inflate desire and short-circuit deliberation. Manufactured-urgency diagnostic.

#### lens:social-proof
- **Foregrounded lens:** `social-proof`
- **Host mode:** `propaganda-audit`
- **Prime prompt:** Audit the social-proof tactic in this ad's '10,000 customers can't be wrong' claim — how others' behavior is used to steer the uncertain buyer.

#### lens:stanley-propaganda
- **Foregrounded lens:** `stanley-propaganda`
- **Host mode:** `propaganda-audit`
- **Prime prompt:** Propaganda audit, Stanley test, on this 2026 campaign ad: 'Real patriots protect our communities. The coastal elites and their open-border allies want you defenseless — only we stand between your family and the chaos they invite.' Supporting vs undermining propaganda — name the democratic ideal it claims and the anti-democratic content it smuggles in, and the flawed belief it activates rather than argues for. Concept-substitution, not-at-issue content.

#### lens:walton-schemes-and-critical-questions
- **Foregrounded lens:** `walton-schemes-and-critical-questions`
- **Host mode:** `coherence-audit`
- **Prime prompt:** Does this argument hold up: 'We should trust the experts on this — the leading authorities agree, so the conclusion stands'? Check the logic and run Walton's argumentation-scheme critical questions for appeal-to-expert-opinion to test whether it actually holds — expertise, consensus, bias, and backing evidence.

#### lens:arrows-impossibility-theorem
- **Foregrounded lens:** `arrows-impossibility-theorem`
- **Host mode:** `boundary-critique`
- **Prime prompt:** Analyze our committee's ranked-vote deadlock as a multi-criteria social-choice problem through Arrow's impossibility theorem — why no voting system over three or more options satisfies all fairness criteria.

#### lens:bounded-rationality
- **Foregrounded lens:** `bounded-rationality`
- **Host mode:** `boundary-critique`
- **Prime prompt:** Analyze our procurement decisions through bounded rationality — how people satisfice under limited time, information, and attention rather than optimize.

#### lens:free-rider-problem
- **Foregrounded lens:** `free-rider-problem`
- **Host mode:** `boundary-critique`
- **Prime prompt:** Our open-source project has thousands of corporate users but only a handful contribute back. Do a boundary critique of the free-rider problem — who benefits without paying, where the system boundary of 'who should contribute' is drawn, and who is wrongly left outside it.

#### lens:tragedy-of-the-commons
- **Foregrounded lens:** `tragedy-of-the-commons`
- **Host mode:** `boundary-critique`
- **Prime prompt:** Who counts as a legitimate claimant on a shared regional aquifer, and where should the system boundary be drawn? Hundreds of farms pump freely, draining it — a tragedy of the commons. Do a boundary critique of whose interests sit inside the boundary, whose are excluded, and how to govern the commons.

#### lens:ulrich-csh-boundary-categories
- **Foregrounded lens:** `ulrich-csh-boundary-categories`
- **Host mode:** `boundary-critique`
- **Prime prompt:** Boundary critique of a city's new 'smart policing' predictive system using Ulrich's critical systems heuristics and its twelve boundary categories — whose interests, expertise, and voices the design includes and whose it excludes from the definitions of 'high-risk area' and 'person of interest'. Sources of motivation, power, knowledge, and legitimacy — who is left out of the boundary judgments?

#### lens:bottlenecks
- **Foregrounded lens:** `bottlenecks`
- **Host mode:** `constraint-mapping`
- **Prime prompt:** Find the bottleneck constraining throughput in our delivery pipeline — the one slowest step that governs the whole system's output.

#### lens:decision-trees
- **Foregrounded lens:** `decision-trees`
- **Host mode:** `decision-architecture`
- **Prime prompt:** Full decision-architecture analysis on a real decision we're facing this quarter — a structured decision document integrating stakeholders, options, and criteria: launching the new product this quarter versus building out two more features first versus a limited regional pilot. Build the decision tree underneath it — options, branch probabilities, payoffs, expected-value rollback. Decision architecture, not just one tree.

#### lens:endowment-effect
- **Foregrounded lens:** `endowment-effect`
- **Host mode:** `decision-architecture`
- **Prime prompt:** This is a big decision and I want the full decision architecture: keep building on our in-house analytics platform or migrate to the off-the-shelf product, taking everything into account — alternatives, what binds each, who's affected, how the leading choice could fail. And apply the 'would we buy our current system today at its true cost?' check so we're not over-valuing it just because we already own it.

#### lens:loss-aversion
- **Foregrounded lens:** `loss-aversion`
- **Host mode:** `decision-architecture`
- **Prime prompt:** Analyze the loss aversion at work when investors hold a losing stock far too long rather than realize the loss — why the pain of the loss outweighs the rational decision to sell.

#### lens:mcdm-methods
- **Foregrounded lens:** `mcdm-methods`
- **Host mode:** `multi-criteria-decision`
- **Prime prompt:** Apply multi-criteria decision methods to choosing a new headquarters city — weight the criteria (cost, talent, taxes) and score the options.

#### lens:prospect-theory
- **Foregrounded lens:** `prospect-theory`
- **Host mode:** `decision-architecture`
- **Prime prompt:** Apply prospect theory to why customers reject a sure small surcharge but accept a risky larger one — outcomes weighed against a reference point, not in absolutes.

#### lens:satisficing
- **Foregrounded lens:** `satisficing`
- **Host mode:** `constraint-mapping`
- **Prime prompt:** Compare our vendor alternatives and map the tradeoffs across cost, reliability, and integration — but through a satisficing lens: we want to stop at the first option that clears a 'good enough' bar on every criterion rather than search for the global optimum. Map the tradeoffs and show when satisficing beats optimizing here.

#### lens:sunk-cost-fallacy
- **Foregrounded lens:** `sunk-cost-fallacy`
- **Host mode:** `decision-architecture`
- **Prime prompt:** Full decision architecture analysis on a real decision we're facing this quarter — choosing among present options: 'keep funding our four-years-and-$12M in-house ERP rebuild, or kill it and switch to the off-the-shelf system.' Structured decision document — stakeholders, options, criteria, and the forward expected value of each path, integrated. Decision architecture — and explicitly keep the money already sunk out of the forward call, separating it from what finishing would cost from here.

#### lens:trade-offs
- **Foregrounded lens:** `trade-offs`
- **Host mode:** `multi-criteria-decision`
- **Prime prompt:** Analyze the core trade-offs in choosing between speed, cost, and quality for our build across these competing criteria — every choice has a cost, so weigh what we give up for what we get.

#### lens:bennett-checkel-process-tracing-tests
- **Foregrounded lens:** `bennett-checkel-process-tracing-tests`
- **Host mode:** `process-tracing`
- **Prime prompt:** Trace the causal process behind our outage using Bennett and Checkel's process-tracing tests — straw-in-the-wind, hoop, smoking-gun, and doubly-decisive evidence.

#### lens:five-whys
- **Foregrounded lens:** `five-whys`
- **Host mode:** `root-cause-analysis`
- **Prime prompt:** Run a five-whys analysis on why the deployment failed — keep asking why until you reach the systemic root cause, not the proximate one.

#### lens:fundamental-attribution-error
- **Foregrounded lens:** `fundamental-attribution-error`
- **Host mode:** `root-cause-analysis`
- **Prime prompt:** What are the root causes of our recurring production outages — we keep blaming the on-call engineer's mistakes, but the same failures recur every month despite repeated fixes. Why does this keep happening? Draw a fishbone separating the systemic causes from the individual — the fundamental attribution error we keep making.

#### lens:pearl-causal-graphs
- **Foregrounded lens:** `pearl-causal-graphs`
- **Host mode:** `causal-dag`
- **Prime prompt:** Does our new onboarding flow actually cause higher retention, or does it just correlate because more motivated users opt in? Draw the causal DAG and apply Pearl's ladder of causation to separate association from intervention — identifying confounders and back-door paths.

#### lens:pearl-do-calculus
- **Foregrounded lens:** `pearl-do-calculus`
- **Host mode:** `causal-dag`
- **Prime prompt:** Causal DAG with Pearl's do-calculus: estimate the effect of our 15% price increase on churn from observational data alone. Draw the DAG with confounders (segment, contract type, seasonality) and mediators, then use do-calculus and the backdoor criterion to decide whether P(churn | do(price)) is identifiable — separating the intervention from mere correlation in the historical data.

#### lens:reward-undermining
- **Foregrounded lens:** `reward-undermining`
- **Host mode:** `systems-dynamics-causal`
- **Prime prompt:** Analyze how paying kids to read undermined their intrinsic love of reading — reward undermining, where external incentives crowd out internal motivation.

#### lens:base-rate-neglect
- **Foregrounded lens:** `base-rate-neglect`
- **Host mode:** `bayesian-hypothesis-network`
- **Prime prompt:** A manager is sure a candidate who 'looks like a founder' will succeed, ignoring that most startups fail. Evaluate the competing hypotheses while correcting for base-rate neglect — how common is success in the reference class before the vivid details.

#### lens:bayesian-reasoning
- **Foregrounded lens:** `bayesian-reasoning`
- **Host mode:** `bayesian-hypothesis-network`
- **Prime prompt:** I got a positive result on a 95%-accurate test for a condition affecting 1 in 1,000 people. Weigh the competing hypotheses — true disease versus false positive — by the diagnosticity of the evidence and the base rate, Bayesian-style, to see which fits best.

#### lens:confirmation-bias
- **Foregrounded lens:** `confirmation-bias`
- **Host mode:** `competing-hypotheses`
- **Prime prompt:** I've already concluded our customer churn is fundamentally a pricing problem, and now I'm reviewing the data. Lay out the competing hypotheses for the churn and stress-test for confirmation bias — where I'd favor evidence that fits the pricing story and discount what cuts against it.

#### lens:differential-diagnosis-schema
- **Foregrounded lens:** `differential-diagnosis-schema`
- **Host mode:** `differential-diagnosis`
- **Prime prompt:** Run a differential diagnosis on our sudden traffic drop — list the candidate causes, then use discriminating tests to rule each in or out.

#### lens:falsifiability
- **Foregrounded lens:** `falsifiability`
- **Host mode:** `competing-hypotheses`
- **Prime prompt:** Lay out the competing hypotheses for this market theory and assess its falsifiability — what observation would disconfirm it, or is it unfalsifiable?

#### lens:representativeness-heuristic
- **Foregrounded lens:** `representativeness-heuristic`
- **Host mode:** `differential-diagnosis`
- **Prime prompt:** Analyze the representativeness heuristic in why we diagnosed the outage as 'probably the database again' because it resembled the last one, ignoring base rates.

#### lens:klein-pre-mortem
- **Foregrounded lens:** `klein-pre-mortem`
- **Host mode:** `pre-mortem-action`
- **Prime prompt:** Pre-mortem this plan: we're launching a redesigned product next quarter — new three-tier pricing page, annual-discount toggle, and a self-serve onboarding flow replacing sales-assisted setup — to our full user base. Imagine it's six months later and the launch failed badly. Run a Klein pre-mortem: work backward from the imagined failure to the preventable causes we can still act on. What killed it?

#### lens:premortem-analysis
- **Foregrounded lens:** `premortem-analysis`
- **Host mode:** `pre-mortem-action`
- **Prime prompt:** Imagine our launch has already failed — run a premortem analysis to surface the failure modes while we can still prevent them.

#### lens:regression-to-mean
- **Foregrounded lens:** `regression-to-mean`
- **Host mode:** `probabilistic-forecasting`
- **Prime prompt:** Our best sales rep just had a record quarter and now we forecast even more. Analyze the regression to the mean — why extreme performance tends to be followed by something closer to average, independent of any intervention.

#### lens:second-order-thinking
- **Foregrounded lens:** `second-order-thinking`
- **Host mode:** `consequences-and-sequel`
- **Prime prompt:** Apply second-order thinking to a rent-control policy — the immediate relief and then the second- and third-order consequences for housing supply.

#### lens:sensemaking
- **Foregrounded lens:** `sensemaking`
- **Host mode:** `scenario-planning`
- **Prime prompt:** Scenario planning for our team facing an ambiguous, fast-moving market shift we can't yet interpret — explore how it might unfold with a 2x2 scenario matrix across the two biggest uncertainties. Treat the exercise as Weickian sensemaking: act into the ambiguity, see what each scenario reveals, and retrospectively build the meaning we need to navigate it.

#### lens:tetlock-superforecasting
- **Foregrounded lens:** `tetlock-superforecasting`
- **Host mode:** `probabilistic-forecasting`
- **Prime prompt:** Give me a calibrated forecast: what is the probability that US headline CPI is below 3% year-over-year by Q4 2026? Use Tetlock-style superforecasting discipline — anchor on the base rate, adjust on indicators, commit to an explicit probability with reasoning.

#### lens:wisdom-of-crowds
- **Foregrounded lens:** `wisdom-of-crowds`
- **Host mode:** `probabilistic-forecasting`
- **Prime prompt:** Analyze when aggregating many independent guesses beats the experts — wisdom of crowds, and the conditions (independence, diversity) it requires.

#### lens:margin-of-safety
- **Foregrounded lens:** `margin-of-safety`
- **Host mode:** `fragility-antifragility-audit`
- **Prime prompt:** Audit the margin of safety in our cash runway — the buffer between what we expect and what we could survive if revenue halved.

#### lens:normal-accident-theory
- **Foregrounded lens:** `normal-accident-theory`
- **Host mode:** `fragility-antifragility-audit`
- **Prime prompt:** Analyze our trading platform through normal accident theory — it is both highly complex and tightly coupled, so analyze where interacting failures could cascade into an accident no one designed.

#### lens:normalization-of-deviance
- **Foregrounded lens:** `normalization-of-deviance`
- **Host mode:** `fragility-antifragility-audit`
- **Prime prompt:** Our ops team keeps skipping a safety check because nothing has gone wrong yet. Analyze this as normalization of deviance — how a tolerated violation becomes the new baseline until it fails.

#### lens:recovery-window
- **Foregrounded lens:** `recovery-window`
- **Host mode:** `fragility-antifragility-audit`
- **Prime prompt:** Analyze the recovery window in our incident response — the time between the first ambiguous warning and irreversible damage, and whether we act inside it.

#### lens:swiss-cheese-model
- **Foregrounded lens:** `swiss-cheese-model`
- **Host mode:** `fragility-antifragility-audit`
- **Prime prompt:** Analyze our outage through the Swiss cheese model — how the holes in monitoring, on-call, and rollback layers lined up to let the failure pass all the way through.

#### lens:taleb-fragility-antifragility
- **Foregrounded lens:** `taleb-fragility-antifragility`
- **Host mode:** `fragility-antifragility-audit`
- **Prime prompt:** Our operations run lean: single-source components, just-in-time inventory, no buffers. Audit this through Taleb's fragility-antifragility lens — where small shocks stay small, where large shocks produce disproportionate convex losses, and whether any part gains from disorder.

#### lens:incentives
- **Foregrounded lens:** `incentives`
- **Host mode:** `stakeholder-mapping`
- **Prime prompt:** Map the stakeholders and their incentives behind why our sales team games the quota system — never ask why people behave a certain way until you see what they are rewarded for.

#### lens:stakeholder-analysis-frameworks
- **Foregrounded lens:** `stakeholder-analysis-frameworks`
- **Host mode:** `stakeholder-mapping`
- **Prime prompt:** Map the stakeholders of our new policy using stakeholder-analysis frameworks — a power/interest grid of who to engage, inform, or watch.

#### lens:framing-effect
- **Foregrounded lens:** `framing-effect`
- **Host mode:** `frame-comparison`
- **Prime prompt:** Compare two framings of the same treatment — '90% survival' versus '10% mortality' — and analyze the framing effect that makes identical facts produce different choices.

#### lens:lakoff-conceptual-metaphor
- **Foregrounded lens:** `lakoff-conceptual-metaphor`
- **Host mode:** `frame-comparison`
- **Prime prompt:** Frame comparison on the immigration debate — two competing conceptual-metaphor frames. The flood frame: 'Immigration is a flood to hold back — a surge at the border, waves of arrivals, a country being swamped.' The newcomer frame: 'Immigrants are newcomers seeking shelter — guests at the door, strivers and contributors, a nation built by immigrants.' Compare both worldviews on their own terms, using conceptual-metaphor analysis to show how each metaphor structures the reasoning and makes different policies feel natural.

#### lens:cappelen-plunkett-conceptual-engineering
- **Foregrounded lens:** `cappelen-plunkett-conceptual-engineering`
- **Host mode:** `conceptual-engineering`
- **Prime prompt:** Conceptual engineering on 'fairness' as used in our hiring policy — it's doing several incompatible jobs at once (equal treatment, equal outcomes, bias-correction, meritocracy) and the inherited concept isn't doing the work it should. Following Cappelen and Plunkett, should we re-engineer what 'fairness' ought to mean for this policy rather than just describe how it's used? What should it mean?

#### lens:map-territory
- **Foregrounded lens:** `map-territory`
- **Host mode:** `conceptual-engineering`
- **Prime prompt:** Analyze the map-territory confusion where our north-star metric became the map mistaken for the territory — clarify how the proxy diverged from the real thing it was meant to represent.

#### lens:system-one-system-two
- **Foregrounded lens:** `system-one-system-two`
- **Host mode:** `deep-clarification`
- **Prime prompt:** Explain in depth how our hiring decisions actually run on fast System 1 intuition versus slow System 2 deliberation — the cognitive mechanics of where the snap pattern-match takes over and where it misfires into bias. I want the depth, Kahneman's dual-process account, not a summary.

#### lens:niches
- **Foregrounded lens:** `niches`
- **Host mode:** `relationship-mapping`
- **Prime prompt:** Relationship map of how the players in our market interact and where each one's competitive niche sits — a dependency graph of who competes with whom, who depends on whom, and the specific conditions and resources where our product outcompetes. What happens to the map when a rival enters our niche?

#### lens:scale
- **Foregrounded lens:** `scale`
- **Host mode:** `relationship-mapping`
- **Prime prompt:** Relationship map of how our system's components interact and which couplings break as we scale from a thousand users to a million — the dependency graph of what affects what, and where the dependencies that hold at small scale snap at large scale.

#### lens:allisons-three-lenses
- **Foregrounded lens:** `allisons-three-lenses`
- **Host mode:** `dialectical-analysis`
- **Prime prompt:** Worldview cartography of the US withdrawal from Afghanistan through Allison's three explanatory paradigms: the rational-actor model, the organizational-process model, and the governmental-politics model. Multi-paradigm map — where do the three lenses cohere on what happened, and where do they irreducibly conflict? Then integrate them into one reading.

#### lens:evolution-natural-selection
- **Foregrounded lens:** `evolution-natural-selection`
- **Host mode:** `synthesis`
- **Prime prompt:** Synthesize how natural selection — variation, selection, and inheritance — explains which startup business models survive in a market.

#### lens:batna
- **Foregrounded lens:** `batna`
- **Host mode:** `interest-mapping`
- **Prime prompt:** I'm a mid-size component supplier renegotiating an exclusive contract with my largest buyer, who keeps hinting they'll switch vendors. Map the underlying interests on both sides and work out my BATNA — my best alternative if this deal collapses — so I know my real leverage.

#### lens:cooperation
- **Foregrounded lens:** `cooperation`
- **Host mode:** `interest-mapping`
- **Prime prompt:** Interest mapping for a repeated negotiation between two rival firms. What does each side really want beneath its positions, and how does mutual cooperation stay stable across repeated rounds despite the temptation to defect? Map the interests and the conditions that make cooperation a stable negotiated outcome.

#### lens:fisher-ury-principled-negotiation
- **Foregrounded lens:** `fisher-ury-principled-negotiation`
- **Host mode:** `interest-mapping`
- **Prime prompt:** Principled negotiation prep for our vendor-contract dispute, full Fisher-Ury treatment: the SaaS vendor wants a 30% renewal increase and a three-year lock-in; we want flexibility and a better rate. Separate the people from the problem, focus on interests not positions, generate options for mutual gain (term length, usage tiers, support SLAs), and find objective criteria for fair pricing. BATNA on both sides.

#### lens:procedural-justice
- **Foregrounded lens:** `procedural-justice`
- **Host mode:** `third-side`
- **Prime prompt:** Analyze why employees accepted an unpopular layoff because the process felt fair — procedural justice, where fairness of process matters as much as the outcome.

#### lens:psychological-safety
- **Foregrounded lens:** `psychological-safety`
- **Host mode:** `third-side`
- **Prime prompt:** Analyze why our team stopped raising concerns, through psychological safety — whether people feel safe to speak up, ask questions, and admit mistakes.

#### lens:ury-third-side
- **Foregrounded lens:** `ury-third-side`
- **Host mode:** `third-side`
- **Prime prompt:** Analyze this community conflict through Ury's Third Side — the surrounding community's role as containers, mediators, and healers of the dispute.

#### lens:circle-of-competence
- **Foregrounded lens:** `circle-of-competence`
- **Host mode:** `domain-induction`
- **Prime prompt:** Map my circle of competence before this biotech investment — where I have genuine earned knowledge and where I only think I do.

#### lens:cynefin-framework
- **Foregrounded lens:** `cynefin-framework`
- **Host mode:** `terrain-mapping`
- **Prime prompt:** I'm trying to get oriented on a problem we can't yet categorize. Walk me through the big picture using the Cynefin framework — is the terrain clear, complicated, complex, or chaotic, and what response does each domain demand? Concept map of the decision terrain.

#### lens:ooda-loop
- **Foregrounded lens:** `ooda-loop`
- **Host mode:** `domain-induction`
- **Prime prompt:** Analyze how a fighter pilot's OODA loop — observe, orient, decide, act — lets them out-tempo a slower opponent, and what 'getting inside the enemy's OODA loop' means here.

#### lens:pareto-principle
- **Foregrounded lens:** `pareto-principle`
- **Host mode:** `quick-orientation`
- **Prime prompt:** Apply the Pareto principle to our support tickets — which 20% of issues cause 80% of the volume, to orient where to focus first.

#### lens:cia-tradecraft-red-team
- **Foregrounded lens:** `cia-tradecraft-red-team`
- **Host mode:** `red-team-advocate`
- **Prime prompt:** Red-team our security assumption that 'no one would target a company our size' using CIA-tradecraft red-team methods — challenge the assumption the way an adversary's analyst would.

#### lens:devils-advocacy
- **Foregrounded lens:** `devils-advocacy`
- **Host mode:** `steelman-construction`
- **Prime prompt:** Steelman the case against our consensus decision to expand into Europe: the room has unanimously agreed to open a London office, hire a regional GM, and localize the product for the EU market on the strength of inbound demand. Play devil's advocate — construct the strongest version of the objection the room is suppressing, best case against, no caveat-padding.

#### lens:groupthink
- **Foregrounded lens:** `groupthink`
- **Host mode:** `red-team-assessment`
- **Prime prompt:** Stress-test this board-approved plan before we commit: our leadership team unanimously signed off on betting the company on a single flagship AI product and sunsetting our three profitable legacy lines by Q4. One cohesive team, a hard deadline, no outside review. Find the holes, rank them by how badly they'd hurt us, and flag any assumption nobody in the room ever challenged. What would a hostile reviewer say?

#### lens:hanlons-razor
- **Foregrounded lens:** `hanlons-razor`
- **Host mode:** `balanced-critique`
- **Prime prompt:** Balanced critique of our vendor's botched data migration that lost three days of customer records: their post-mortem cites an untested backup script and a skipped staging run. Strengths and weaknesses at equal depth, neutral read, no advocacy — and as you weigh the failures, judge whether each points to ordinary incompetence or constraint rather than deliberate negligence.

#### lens:hindsight-bias
- **Foregrounded lens:** `hindsight-bias`
- **Host mode:** `benefits-analysis`
- **Prime prompt:** Analyze the hindsight bias in our incident review — how 'we should have seen it coming' rewrites what was actually knowable beforehand.

#### lens:narrative-instinct
- **Foregrounded lens:** `narrative-instinct`
- **Host mode:** `balanced-critique`
- **Prime prompt:** Analyze the narrative instinct in how this post-mortem turned a messy accident into a clean story with a villain — the human compulsion to build a coherent narrative.

#### lens:occams-razor
- **Foregrounded lens:** `occams-razor`
- **Host mode:** `balanced-critique`
- **Prime prompt:** Apply Occam's razor to these competing explanations for our outage — prefer the one that needs the fewest assumptions to fit the evidence.

#### lens:emergence
- **Foregrounded lens:** `emergence`
- **Host mode:** `mechanism-understanding`
- **Prime prompt:** How does a traffic jam actually work under the hood — the structural explanation of how each driver's simple local rules (follow, brake, change lanes) produce the emergent jam, a global pattern the individual parts don't have? Parts to behavior, the mechanism of emergence.

#### lens:first-principles
- **Foregrounded lens:** `first-principles`
- **Host mode:** `mechanism-understanding`
- **Prime prompt:** Reason from first principles about whether we actually need a native app — decompose the requirement to fundamentals and rebuild, ignoring what competitors do.

#### lens:creative-destruction
- **Foregrounded lens:** `creative-destruction`
- **Host mode:** `market-dynamics`
- **Prime prompt:** Cheap AI transcription is wiping out the traditional medical-transcription industry. Analyze this through Schumpeter's creative destruction — how the new technology destroys the incumbent's rent and what the market looks like afterward.

#### lens:critical-mass
- **Foregrounded lens:** `critical-mass`
- **Host mode:** `market-dynamics`
- **Prime prompt:** Our two-sided marketplace has 5,000 buyers but only 400 sellers and growth is sluggish. Analyze the critical-mass dynamics — where the network-effects tipping point is and what it takes to reach self-sustaining growth.

#### lens:diminishing-returns
- **Foregrounded lens:** `diminishing-returns`
- **Host mode:** `market-dynamics`
- **Prime prompt:** We keep adding engineers to the same product squad and shipping speed has barely moved. Analyze the diminishing returns — where the marginal payoff of another hire bends — and what that means for our next allocation in this market for talent.

#### lens:equilibrium
- **Foregrounded lens:** `equilibrium`
- **Host mode:** `market-dynamics`
- **Prime prompt:** Analyze the market equilibrium in our two-sided marketplace — the state where buyer demand and seller supply balance, and what shocks knock it out of balance.

#### lens:feedback-loops
- **Foregrounded lens:** `feedback-loops`
- **Host mode:** `systems-dynamics-structural`
- **Prime prompt:** Draw the feedback structure of our customer-success system — the reinforcing and balancing loops and the stocks and flows linking onboarding quality, churn, and referrals — and show which loops dominate over time.

#### lens:greshams-law
- **Foregrounded lens:** `greshams-law`
- **Host mode:** `market-dynamics`
- **Prime prompt:** On our freelance marketplace the cheap low-quality providers keep crowding out the skilled ones, and buyers can't tell them apart before hiring. Analyze this as a Gresham's law dynamic — how the bad drives out the good — and what it does to the market over time.

#### lens:leverage
- **Foregrounded lens:** `leverage`
- **Host mode:** `systems-dynamics-structural`
- **Prime prompt:** Draw the feedback structure of our customer-retention system — the stocks (active users, at-risk users, churned), the flows (signups, churn, reactivation), and the reinforcing and balancing loops. Then identify the Meadows leverage points — the few places in that feedback structure where a small, well-placed intervention shifts the whole system. Structural picture first.

#### lens:practical-drift
- **Foregrounded lens:** `practical-drift`
- **Host mode:** `process-mapping`
- **Prime prompt:** Analyze how our deployment process drifted from the written procedure to 'how we actually do it' — practical drift, and the risk it accumulates.

#### lens:red-queen-effect
- **Foregrounded lens:** `red-queen-effect`
- **Host mode:** `market-dynamics`
- **Prime prompt:** Analyze the market dynamics in our retail category as a Red Queen effect: every retailer matches the others' loyalty programs and discounts, nobody gains share, and all just spend more to stay even — competitive coevolution that runs to stand still.

#### lens:supply-demand
- **Foregrounded lens:** `supply-demand`
- **Host mode:** `market-dynamics`
- **Prime prompt:** Our city is about to approve 10,000 new apartments over three years. Walk me through the supply and demand dynamics — what happens to rents in the short run versus the long run as the new supply comes online?

#### lens:adverse-selection
- **Foregrounded lens:** `adverse-selection`
- **Host mode:** `mechanism-design`
- **Prime prompt:** Our new pay-monthly device warranty is attracting mostly customers whose phones are already failing, and we're losing money. The customers know their device's condition and we don't. Analyze the adverse selection — who self-selects in, who exits — and how screening could separate the risk types.

#### lens:asymmetric-warfare
- **Foregrounded lens:** `asymmetric-warfare`
- **Host mode:** `strategic-interaction`
- **Prime prompt:** Game theory of an asymmetric conflict: what's the weaker player's best move against a dominant but rigid incumbent — refusing to fight on the incumbent's terms and attacking where the big player is strong but inflexible? Payoff matrix, the asymmetric-warfare dynamics, deterrence and where the incumbent's commitments aren't credible.

#### lens:backward-induction
- **Foregrounded lens:** `backward-induction`
- **Host mode:** `strategic-interaction`
- **Prime prompt:** Game theory: solve this sequential bargaining game by backward induction. Two players split $100 over three alternating-offer rounds; the pie shrinks 20% each round an offer is rejected. Start from the final round and reason back to the first move to find each player's optimal strategy. Payoff at each node, subgame-perfect path.

#### lens:brinkmanship
- **Foregrounded lens:** `brinkmanship`
- **Host mode:** `strategic-interaction`
- **Prime prompt:** Two nuclear powers escalate a border standoff toward the brink. Using game theory, analyze the brinkmanship and deterrence: how going to the edge of catastrophe makes the threat credible, what each side expects the other to do, and where it tips into disaster.

#### lens:moral-hazard
- **Foregrounded lens:** `moral-hazard`
- **Host mode:** `mechanism-design`
- **Prime prompt:** After we began fully reimbursing field reps for any client dinner with no cap, spending exploded — and we can't observe each meal. Analyze the moral hazard, where the reps' hidden action diverges from our interest once they're shielded from the cost, and how to redesign the incentives.

#### lens:mutually-assured-destruction
- **Foregrounded lens:** `mutually-assured-destruction`
- **Host mode:** `strategic-interaction`
- **Prime prompt:** Analyze a price war between two dominant platforms as mutually assured destruction — each can destroy the other's margins, so the credible threat of mutual ruin enforces an uneasy peace. Use game theory.

#### lens:nash-equilibrium
- **Foregrounded lens:** `nash-equilibrium`
- **Host mode:** `strategic-interaction`
- **Prime prompt:** Two airlines, Delta and United, compete on one route and each must decide simultaneously every quarter whether to price High or Low. Payoffs: both High, each earns 100; one prices Low while the other holds High, the discounter earns 140 and the rival 40; both Low, each earns 60. Moves are simultaneous, no communication, repeated each quarter. Find the Nash equilibrium of this game.

#### lens:principal-agent-problem
- **Foregrounded lens:** `principal-agent-problem`
- **Host mode:** `mechanism-design`
- **Prime prompt:** Our sales reps are paid on bookings, but they privately control which deals they chase and how they discount, and we can't see their effort. Analyze the principal-agent problem — the hidden action and misaligned incentives — and how the contract could be redesigned.

#### lens:prisoners-dilemma
- **Foregrounded lens:** `prisoners-dilemma`
- **Host mode:** `strategic-interaction`
- **Prime prompt:** Two competitors each decide whether to cut prices, where mutual restraint is best for both but each is tempted to defect. Analyze this as a prisoner's dilemma — why rational self-interest produces a worse outcome for both. Use game theory.

#### lens:schelling-point
- **Foregrounded lens:** `schelling-point`
- **Host mode:** `strategic-interaction`
- **Prime prompt:** Model this as a coordination game: two rival firms independently pick a product standard with no communication, and both gain only if they match. Using game theory, find the focal Schelling point they converge on.

#### lens:signaling
- **Foregrounded lens:** `signaling`
- **Host mode:** `strategic-interaction`
- **Prime prompt:** Game theory of costly signaling: model why a company burns millions on a Super Bowl ad that says nothing about the product. Set up the payoff structure — the ad is a costly, hard-to-fake action that credibly signals private information about quality to buyers who can't otherwise verify it. Separating versus pooling equilibrium, why cheap talk fails.

#### lens:tit-for-tat
- **Foregrounded lens:** `tit-for-tat`
- **Host mode:** `strategic-interaction`
- **Prime prompt:** Two competitors repeatedly choose whether to honor or break a price truce. Using game theory on this repeated game, analyze why tit-for-tat — cooperate first, then mirror the rival's last move — is the best move.

#### lens:winners-curse
- **Foregrounded lens:** `winners-curse`
- **Host mode:** `mechanism-design`
- **Prime prompt:** We keep winning competitive acquisition auctions and then finding the target was worth less than we paid — each bidder privately estimates value and the highest estimate wins. Analyze the winner's curse — why winning is itself bad news about value — and what it implies for how we bid.

#### lens:alexander-pattern-language
- **Foregrounded lens:** `alexander-pattern-language`
- **Host mode:** `place-reading-genius-loci`
- **Prime prompt:** Read this town square through Alexander's pattern language — which patterns ('small public squares', 'building edges', 'activity pockets') are present or violated, and how they shape whether people gather.

#### lens:appleton-prospect-refuge
- **Foregrounded lens:** `appleton-prospect-refuge`
- **Host mode:** `place-reading-genius-loci`
- **Prime prompt:** Read this café's seating through Appleton's prospect-refuge theory — why the corner booths with a view of the door fill first and the exposed center tables last.

#### lens:arnheim-compositional-forces
- **Foregrounded lens:** `arnheim-compositional-forces`
- **Host mode:** `compositional-dynamics`
- **Prime prompt:** Here is our new landing-page hero layout — a big headline upper-left, an image lower-right, a button dead center. Read its compositional dynamics through Arnheim's structural skeleton and visual forces: where the axes and centers pull the eye, and where the composition is balanced or strained.

#### lens:bachelard-topoanalysis
- **Foregrounded lens:** `bachelard-topoanalysis`
- **Host mode:** `place-reading-genius-loci`
- **Prime prompt:** Read the intimate spaces of a childhood home through Bachelard's topoanalysis — how corners, attics, and drawers hold psychological resonance, not just square footage.

#### lens:bertin-visual-variables
- **Foregrounded lens:** `bertin-visual-variables`
- **Host mode:** `information-density`
- **Prime prompt:** Critique this data map with a Bertin visual-variables audit: a US county map encoding three variables at once — population density as color saturation (pale yellow to deep red), median income as circle size, and unemployment as circle hue (green to purple), all stacked on the same counties. Through Bertin's retinal variables, which variable rides which retinal variable, where the encodings collide, and where the encoding fights perception. Prescriptive recommendations.

#### lens:cleveland-mcgill-perceptual-tasks
- **Foregrounded lens:** `cleveland-mcgill-perceptual-tasks`
- **Host mode:** `information-density`
- **Prime prompt:** Critique this chart with a Cleveland-McGill perceptual-task audit: a quarterly market-share figure drawn as a 3-D exploded pie of six wedges, value labels hidden behind the tilt, rainbow legend off to the side. Analyze the information density through Cleveland and McGill's perceptual-task ranking — whether readers must judge position, length, angle, or area, and how that predicts reading accuracy. Prescriptive recommendations.

#### lens:gestalt-grouping-principles
- **Foregrounded lens:** `gestalt-grouping-principles`
- **Host mode:** `compositional-dynamics`
- **Prime prompt:** Here is the layout of our analytics dashboard — twelve tiles, three accent colors, uneven spacing. Read its compositional dynamics through the Gestalt grouping principles: what proximity, similarity, and figure-ground organization make the eye bind, and where grouping fights meaning.

#### lens:japanese-aesthetics-catalog
- **Foregrounded lens:** `japanese-aesthetics-catalog`
- **Host mode:** `ma-reading`
- **Prime prompt:** Read the use of Ma — negative space and interval — in this minimalist Japanese garden, where the empty raked gravel between the rocks carries as much meaning as the rocks themselves.

#### lens:kaplan-attention-restoration
- **Foregrounded lens:** `kaplan-attention-restoration`
- **Host mode:** `place-reading-genius-loci`
- **Prime prompt:** Read this hospital garden as a place through Kaplan's attention-restoration theory — the soft fascination, being-away, extent, and compatibility that let depleted directed attention recover.

#### lens:lynch-image-of-the-city
- **Foregrounded lens:** `lynch-image-of-the-city`
- **Host mode:** `place-reading-genius-loci`
- **Prime prompt:** Read this downtown's legibility through Lynch's image of the city — its paths, edges, districts, nodes, and landmarks, and where people get lost.

#### lens:norberg-schulz-genius-loci
- **Foregrounded lens:** `norberg-schulz-genius-loci`
- **Host mode:** `place-reading-genius-loci`
- **Prime prompt:** Read the genius loci — the spirit of place — of an old European market square through Norberg-Schulz, the qualitative atmosphere that decomposing it into 'a square plus buildings' misses.

#### lens:tufte-data-ink-chartjunk
- **Foregrounded lens:** `tufte-data-ink-chartjunk`
- **Host mode:** `information-density`
- **Prime prompt:** Critique this report's chart with a Tufte data-ink-ratio audit: a quarterly revenue dashboard with 3-D bars, drop shadows, a rainbow gradient fill on each bar, gridlines every $1k, and a logo watermark behind the plot. Through Tufte's data-ink ratio — how much ink encodes data versus decoration, and what chartjunk to strip to raise the information density. Prescriptive recommendations.

**END OF ORA PERFORMANCE CAPTURE REPRODUCTION FRAMEWORK v1.0**