# Hand-authored text-only development tasks Catalogue version 1.1. These twelve tasks are hand-authored development assets for the [proposed protocol](../03-experimental-protocol.md), with four tasks in each family. Twelve is an **unfrozen development default** matching the proposed pilot allocation, not a registered sample-size or power decision. These are not held-out tests, model outputs, or experimental results. No task has been executed with a worker or planner. The three original scenario packets and their individual assessor guides remain unchanged. | Task | Family | Participant packet | Assessor-only guide | |---|---|---|---| | DEV-SYN-01 | Source-grounded synthesis | [Archive decision brief](public/DEV-SYN-01-archive-brief.md) | [Scoring and supported answers](assessor-only/DEV-SYN-01-guide.md) | | DEV-SYN-02 | Source-grounded synthesis | [Transit survey evidence audit](public/DEV-SYN-02-transit-survey.md) | [Counts, denominators, and claim limits](assessor-only/DEV-SYN-02-guide.md) | | DEV-SYN-03 | Source-grounded synthesis | [Mural attribution dossier](public/DEV-SYN-03-mural-attribution.md) | [Attribution and chronology key](assessor-only/DEV-SYN-03-guide.md) | | DEV-SYN-04 | Source-grounded synthesis | [Material-flow claims review](public/DEV-SYN-04-material-flow.md) | [Mass balance and claim limits](assessor-only/DEV-SYN-04-guide.md) | | DEV-CON-01 | Constraint-heavy documents | [Equipment-lending operating procedure](public/DEV-CON-01-lending-procedure.md) | [Scoring and acceptable alternatives](assessor-only/DEV-CON-01-guide.md) | | DEV-CON-02 | Constraint-heavy documents | [Blind contest administration](public/DEV-CON-02-blind-contest.md) | [Eligibility, clocks, and appeals key](assessor-only/DEV-CON-02-guide.md) | | DEV-CON-03 | Constraint-heavy documents | [Oral-history transcript access](public/DEV-CON-03-transcript-access.md) | [Consent scope and revocation key](assessor-only/DEV-CON-03-guide.md) | | DEV-CON-04 | Constraint-heavy documents | [Game-festival token desk](public/DEV-CON-04-token-desk.md) | [Expiry, duplicate, and correction key](assessor-only/DEV-CON-04-guide.md) | | DEV-PLAN-01 | Structured planning and analysis | [Exhibit dispatch resource plan](public/DEV-PLAN-01-exhibit-dispatch.md) | [Feasibility and arithmetic key](assessor-only/DEV-PLAN-01-guide.md) | | DEV-PLAN-02 | Structured planning and analysis | [Learning-centre rotations](public/DEV-PLAN-02-learning-rotations.md) | [Infeasibility bound and authorized extension](assessor-only/DEV-PLAN-02-guide.md) | | DEV-PLAN-03 | Structured planning and analysis | [Dye-studio water allocation](public/DEV-PLAN-03-water-allocation.md) | [Grade-specific balances and capacity key](assessor-only/DEV-PLAN-03-guide.md) | | DEV-PLAN-04 | Structured planning and analysis | [Short-film screening programme](public/DEV-PLAN-04-screening-programme.md) | [Programme and swap feasibility key](assessor-only/DEV-PLAN-04-guide.md) | ## Provenance and reuse All organizations, people, records, quantities, prices, policies, and calendar windows in these packets are fictional and authored directly for this development set. There is no external dataset, real vendor quotation, empirical pilot observation, or third-party source behind them. Numbers described as historical samples are fictional facts inside the task, not research results. Each individual packet and its matching guide are version 1.0 and immutable: revise under a new packet/guide version if development work later exposes ambiguity. Catalogue version 1.1 adds nine scenarios without changing the original three. Inspection found an existing complete illustration in manuscript Appendix A, prompt sketches in [05-planner-prompts.md](../05-planner-prompts.md), and the historical Launchboard software seed in `Mac/checks/closed-loop/seed.json`. This set reuses the completed task/requirement/source-location convention and the protocol's three task families and scoring approach. It preserves those existing assets. It does not recycle the workshop scenario, its initial artifact or planner response, or the software fixture as new text tasks. No participant deliverables, seeded defects, planner responses, or successful trajectories are included here. The expansion varies the reasoning problem rather than just renaming an organization: survey aggregation and sampling, historical attribution, mass-flow publicity claims, blind judging and appeals, segment-level consent, idempotent credit handling, room-slot infeasibility, grade-specific stock allocation, and programme selection. The [assessor-only expansion review](assessor-only/expansion-review.md) records manual arithmetic, feasibility, source-sufficiency, and mapping checks. It is an authoring review, not a participant trial. ## Participant and assessor separation Each file in `public/` is a self-contained immutable assignment with all public requirements and a complete fixed source packet. Copy only the selected public file into a future worker/planner context, alongside the frozen study instructions. Source and requirement IDs are scoped to that task. No browsing, connected services, code changes, or outside factual assumptions are needed. The files in `assessor-only/` contain criterion mappings and supported answers. They must remain outside participant context and be used only under the protocol's development/assessment separation. Folder naming is organizational separation, not an enforced access-control mechanism; a future runner would need to restrict supplied files. See [assessor conventions](assessor-only/README.md). The public packet defines every scored obligation; assessor guides introduce none. This catalogue now fills the proposed twelve-task development allocation, but neither that count nor the composition is registered or experimentally frozen. Later development may motivate a new version or a different sample-size decision. Once inspected for development, these scenarios and their close variants must not be relabeled as untouched validation or test tasks. No participant outcomes, model-compliance evidence, or performance estimates are provided.