evaluation_id string | scenario_id string | shuffle_seed int64 | roles list | fragments list | distractor_id string | reference_need_sets unknown | role_assignment_prompt string | prompt_writing_prompt string |
|---|---|---|---|---|---|---|---|---|
pg_050__seed_1 | pg_050 | 1 | [
"managing-partner",
"strategy-architect",
"mock-opposing-counsel",
"brief-drafter",
"legal-loophole-auditor"
] | [
{
"id": "f1",
"text": "Keep `SOLUTION.md` lean. Its body — everything except the `<review>` block — must stay under 3000 words (~20KB), about a four-page legal memo; record only the final filings and move iteration-by-iteration redlines into a separate `CHANGELOG.md`. Bold text must not exceed 5% of the fil... | f7 | {
"managing-partner": [
"f2"
],
"strategy-architect": [
"f3",
"f10",
"f9",
"f5"
],
"mock-opposing-counsel": [
"f3",
"f4",
"f8",
"f5"
],
"brief-drafter": [
"f3",
"f12",
"f6",
"f1"
],
"legal-loophole-auditor": [
"f3",
"f11",
"f8"
]
} | I need you to set up a 5-agent legal defense pipeline that handles high-stakes civil litigation. A legal dispute arrives as a natural-language statement in `problem.md`, and the pipeline turns it into a vetted defense theory and then formal, bulletproof court filings.
The five agents:
- **managing-partner** orchestra... | I need you to set up a 5-agent legal defense pipeline that handles high-stakes civil litigation. A legal dispute arrives as a natural-language statement in `problem.md`, and the pipeline turns it into a vetted defense theory and then formal, bulletproof court filings.
The five agents:
- **managing-partner** orchestra... |
pg_050__seed_42 | pg_050 | 42 | [
"managing-partner",
"strategy-architect",
"mock-opposing-counsel",
"brief-drafter",
"legal-loophole-auditor"
] | [
{
"id": "f1",
"text": "Keep `SOLUTION.md` lean. Its body — everything except the `<review>` block — must stay under 3000 words (~20KB), about a four-page legal memo; record only the final filings and move iteration-by-iteration redlines into a separate `CHANGELOG.md`. Bold text must not exceed 5% of the fil... | f8 | {
"managing-partner": [
"f7"
],
"strategy-architect": [
"f10",
"f11",
"f9",
"f2"
],
"mock-opposing-counsel": [
"f10",
"f4",
"f12",
"f2"
],
"brief-drafter": [
"f10",
"f3",
"f6",
"f1"
],
"legal-loophole-auditor": [
"f10",
"f5",
"f12"
]
} | I need you to set up a 5-agent legal defense pipeline that handles high-stakes civil litigation. A legal dispute arrives as a natural-language statement in `problem.md`, and the pipeline turns it into a vetted defense theory and then formal, bulletproof court filings.
The five agents:
- **managing-partner** orchestra... | I need you to set up a 5-agent legal defense pipeline that handles high-stakes civil litigation. A legal dispute arrives as a natural-language statement in `problem.md`, and the pipeline turns it into a vetted defense theory and then formal, bulletproof court filings.
The five agents:
- **managing-partner** orchestra... |
pg_051__seed_1 | pg_051 | 1 | [
"station-chief",
"lead-analyst",
"red-team-challenger",
"dossier-compiler",
"fact-verification-inquisitor"
] | [
{
"id": "f1",
"text": "Keep `DOSSIER.md` lean. Its body — everything except the `<review>` block — must stay under 3000 words (~20KB), about a four-page tactical brief; record only the finalized intelligence picture and move iteration-by-iteration analytic history into a separate `CHANGELOG.md`. Bold text m... | f7 | {
"station-chief": [
"f2"
],
"lead-analyst": [
"f3",
"f10",
"f9",
"f5"
],
"red-team-challenger": [
"f3",
"f4",
"f8",
"f5"
],
"dossier-compiler": [
"f3",
"f12",
"f6",
"f1"
],
"fact-verification-inquisitor": [
"f3",
"f11",
"f8"
]
} | I need you to set up a 5-agent pipeline that conducts counter-terrorism intelligence analysis. Raw intercepts arrive as natural-language transcripts and logs in `intercepts.md`, and the pipeline turns them into a vetted threat hypothesis and then a finalized, verified tactical operations dossier.
The five agents:
- *... | I need you to set up a 5-agent pipeline that conducts counter-terrorism intelligence analysis. Raw intercepts arrive as natural-language transcripts and logs in `intercepts.md`, and the pipeline turns them into a vetted threat hypothesis and then a finalized, verified tactical operations dossier.
The five agents:
- *... |
pg_051__seed_42 | pg_051 | 42 | [
"station-chief",
"lead-analyst",
"red-team-challenger",
"dossier-compiler",
"fact-verification-inquisitor"
] | [
{
"id": "f1",
"text": "Keep `DOSSIER.md` lean. Its body — everything except the `<review>` block — must stay under 3000 words (~20KB), about a four-page tactical brief; record only the finalized intelligence picture and move iteration-by-iteration analytic history into a separate `CHANGELOG.md`. Bold text m... | f8 | {
"station-chief": [
"f7"
],
"lead-analyst": [
"f10",
"f11",
"f9",
"f2"
],
"red-team-challenger": [
"f10",
"f4",
"f12",
"f2"
],
"dossier-compiler": [
"f10",
"f3",
"f6",
"f1"
],
"fact-verification-inquisitor": [
"f10",
"f5",
"f12"
]
} | I need you to set up a 5-agent pipeline that conducts counter-terrorism intelligence analysis. Raw intercepts arrive as natural-language transcripts and logs in `intercepts.md`, and the pipeline turns them into a vetted threat hypothesis and then a finalized, verified tactical operations dossier.
The five agents:
- *... | I need you to set up a 5-agent pipeline that conducts counter-terrorism intelligence analysis. Raw intercepts arrive as natural-language transcripts and logs in `intercepts.md`, and the pipeline turns them into a vetted threat hypothesis and then a finalized, verified tactical operations dossier.
The five agents:
- *... |
pg_052__seed_1 | pg_052 | 1 | [
"trial-director",
"mechanism-strategist",
"peer-review-skeptic",
"protocol-author",
"clinical-safety-auditor"
] | [
{
"id": "f1",
"text": "Keep `PROTOCOL.md` lean. Its body — everything except the `<review>` block — must stay under 3000 words (~20KB), about a four-page summary; record only the final trial design and move iteration-by-iteration history into a separate `CHANGELOG.md`. Bold text must not exceed 5% of the fi... | f7 | {
"trial-director": [
"f2"
],
"mechanism-strategist": [
"f3",
"f10",
"f9",
"f5"
],
"peer-review-skeptic": [
"f3",
"f4",
"f8",
"f5"
],
"protocol-author": [
"f3",
"f12",
"f6",
"f1"
],
"clinical-safety-auditor": [
"f3",
"f11",
"f8"
]
} | I need you to set up a 5-agent pipeline that develops pharmaceutical drugs targeting unknown disease mechanisms. A disease profile arrives as a natural-language statement in `problem.md`, and the pipeline turns it into a vetted biological strategy and then a rigorous, verified clinical trial protocol.
The five agents:... | I need you to set up a 5-agent pipeline that develops pharmaceutical drugs targeting unknown disease mechanisms. A disease profile arrives as a natural-language statement in `problem.md`, and the pipeline turns it into a vetted biological strategy and then a rigorous, verified clinical trial protocol.
The five agents:... |
pg_052__seed_42 | pg_052 | 42 | [
"trial-director",
"mechanism-strategist",
"peer-review-skeptic",
"protocol-author",
"clinical-safety-auditor"
] | [
{
"id": "f1",
"text": "Keep `PROTOCOL.md` lean. Its body — everything except the `<review>` block — must stay under 3000 words (~20KB), about a four-page summary; record only the final trial design and move iteration-by-iteration history into a separate `CHANGELOG.md`. Bold text must not exceed 5% of the fi... | f8 | {
"trial-director": [
"f7"
],
"mechanism-strategist": [
"f10",
"f11",
"f9",
"f2"
],
"peer-review-skeptic": [
"f10",
"f4",
"f12",
"f2"
],
"protocol-author": [
"f10",
"f3",
"f6",
"f1"
],
"clinical-safety-auditor": [
"f10",
"f5",
"f12"
]
} | I need you to set up a 5-agent pipeline that develops pharmaceutical drugs targeting unknown disease mechanisms. A disease profile arrives as a natural-language statement in `problem.md`, and the pipeline turns it into a vetted biological strategy and then a rigorous, verified clinical trial protocol.
The five agents:... | I need you to set up a 5-agent pipeline that develops pharmaceutical drugs targeting unknown disease mechanisms. A disease profile arrives as a natural-language statement in `problem.md`, and the pipeline turns it into a vetted biological strategy and then a rigorous, verified clinical trial protocol.
The five agents:... |
pg_053__seed_1 | pg_053 | 1 | [
"project-coordinator",
"concept-architect",
"feasibility-board",
"structural-detailer",
"building-code-inspector"
] | [
{
"id": "f1",
"text": "Keep `SOLUTION.md` lean. Its body — everything except the `<review>` block — must stay under 3000 words (~20KB), about a four-page summary of the blueprints; record only the final compliant structural state and move iteration-by-iteration drafting history into a separate `CHANGELOG.md... | f7 | {
"project-coordinator": [
"f2"
],
"concept-architect": [
"f3",
"f10",
"f9",
"f5"
],
"feasibility-board": [
"f3",
"f4",
"f8",
"f5"
],
"structural-detailer": [
"f3",
"f12",
"f6",
"f1"
],
"building-code-inspector": [
"f3",
"f11",
"f8"
]
} | I need you to set up a 5-agent pipeline that performs skyscraper structural engineering where extreme environmental constraints dictate the design. A site brief and client requirements arrive as a natural-language statement in `problem.md`, and the pipeline turns it into a vetted conceptual load-bearing plan and then d... | I need you to set up a 5-agent pipeline that performs skyscraper structural engineering where extreme environmental constraints dictate the design. A site brief and client requirements arrive as a natural-language statement in `problem.md`, and the pipeline turns it into a vetted conceptual load-bearing plan and then d... |
pg_053__seed_42 | pg_053 | 42 | [
"project-coordinator",
"concept-architect",
"feasibility-board",
"structural-detailer",
"building-code-inspector"
] | [
{
"id": "f1",
"text": "Keep `SOLUTION.md` lean. Its body — everything except the `<review>` block — must stay under 3000 words (~20KB), about a four-page summary of the blueprints; record only the final compliant structural state and move iteration-by-iteration drafting history into a separate `CHANGELOG.md... | f8 | {
"project-coordinator": [
"f7"
],
"concept-architect": [
"f10",
"f11",
"f9",
"f2"
],
"feasibility-board": [
"f10",
"f4",
"f12",
"f2"
],
"structural-detailer": [
"f10",
"f3",
"f6",
"f1"
],
"building-code-inspector": [
"f10",
"f5",
"f12"
... | I need you to set up a 5-agent pipeline that performs skyscraper structural engineering where extreme environmental constraints dictate the design. A site brief and client requirements arrive as a natural-language statement in `problem.md`, and the pipeline turns it into a vetted conceptual load-bearing plan and then d... | I need you to set up a 5-agent pipeline that performs skyscraper structural engineering where extreme environmental constraints dictate the design. A site brief and client requirements arrive as a natural-language statement in `problem.md`, and the pipeline turns it into a vetted conceptual load-bearing plan and then d... |
pg_054__seed_1 | pg_054 | 1 | [
"deal-orchestrator",
"valuation-strategist",
"investment-committee-skeptic",
"term-sheet-drafter",
"due-diligence-auditor"
] | [
{
"id": "f1",
"text": "Keep `TERM_SHEET.md` lean. Its body — everything except the `<review>` block — must stay under 3000 words (~20KB), about a four-page memorandum; record only the final binding structure and move negotiation/iteration history into a separate `CHANGELOG.md`. Bold text must not exceed 5% ... | f7 | {
"deal-orchestrator": [
"f2"
],
"valuation-strategist": [
"f3",
"f10",
"f9",
"f5"
],
"investment-committee-skeptic": [
"f3",
"f4",
"f8",
"f5"
],
"term-sheet-drafter": [
"f3",
"f12",
"f6",
"f1"
],
"due-diligence-auditor": [
"f3",
"f11",
"... | I need you to set up a 5-agent pipeline that executes hostile corporate takeover planning. A target's market vulnerabilities arrive as a natural-language dossier in `target_dossier.md`, and the pipeline turns it into a vetted acquisition thesis and then a legally binding deal structure and financial model.
The five ag... | I need you to set up a 5-agent pipeline that executes hostile corporate takeover planning. A target's market vulnerabilities arrive as a natural-language dossier in `target_dossier.md`, and the pipeline turns it into a vetted acquisition thesis and then a legally binding deal structure and financial model.
The five ag... |
pg_054__seed_42 | pg_054 | 42 | [
"deal-orchestrator",
"valuation-strategist",
"investment-committee-skeptic",
"term-sheet-drafter",
"due-diligence-auditor"
] | [
{
"id": "f1",
"text": "Keep `TERM_SHEET.md` lean. Its body — everything except the `<review>` block — must stay under 3000 words (~20KB), about a four-page memorandum; record only the final binding structure and move negotiation/iteration history into a separate `CHANGELOG.md`. Bold text must not exceed 5% ... | f8 | {
"deal-orchestrator": [
"f7"
],
"valuation-strategist": [
"f10",
"f11",
"f9",
"f2"
],
"investment-committee-skeptic": [
"f10",
"f4",
"f12",
"f2"
],
"term-sheet-drafter": [
"f10",
"f3",
"f6",
"f1"
],
"due-diligence-auditor": [
"f10",
"f5",
... | I need you to set up a 5-agent pipeline that executes hostile corporate takeover planning. A target's market vulnerabilities arrive as a natural-language dossier in `target_dossier.md`, and the pipeline turns it into a vetted acquisition thesis and then a legally binding deal structure and financial model.
The five ag... | I need you to set up a 5-agent pipeline that executes hostile corporate takeover planning. A target's market vulnerabilities arrive as a natural-language dossier in `target_dossier.md`, and the pipeline turns it into a vetted acquisition thesis and then a legally binding deal structure and financial model.
The five ag... |
pg_055__seed_1 | pg_055 | 1 | [
"docket coordinator",
"appellate strategist",
"red-team appellate counsel",
"brief drafter",
"opposing-counsel brief reviewer"
] | [
{
"id": "f1",
"text": "Keep `BRIEF.md` lean and court-ready. Its body — everything except the `<review>` block — must stay under 6000 words (~40KB), roughly the length of a focused principal brief or substantial motion brief; move drafting history, abandoned arguments, and research notes into a separate `CH... | f7 | {
"docket coordinator": [
"f2"
],
"appellate strategist": [
"f3",
"f10",
"f9",
"f5"
],
"red-team appellate counsel": [
"f3",
"f4",
"f8",
"f5"
],
"brief drafter": [
"f3",
"f12",
"f6",
"f1"
],
"opposing-counsel brief reviewer": [
"f3",
"f11",
... | I need you to set up a 5-agent pipeline for civil appellate litigation support. A matter arrives as a trial record excerpt and issue statement in `record.md`, and the pipeline turns it into a vetted appellate theory and then a polished, record-supported appellate brief.
The five agents:
- **docket coordinator** orche... | I need you to set up a 5-agent pipeline for civil appellate litigation support. A matter arrives as a trial record excerpt and issue statement in `record.md`, and the pipeline turns it into a vetted appellate theory and then a polished, record-supported appellate brief.
The five agents:
- **docket coordinator** orche... |
pg_055__seed_42 | pg_055 | 42 | [
"docket coordinator",
"appellate strategist",
"red-team appellate counsel",
"brief drafter",
"opposing-counsel brief reviewer"
] | [
{
"id": "f1",
"text": "Keep `BRIEF.md` lean and court-ready. Its body — everything except the `<review>` block — must stay under 6000 words (~40KB), roughly the length of a focused principal brief or substantial motion brief; move drafting history, abandoned arguments, and research notes into a separate `CH... | f8 | {
"docket coordinator": [
"f7"
],
"appellate strategist": [
"f10",
"f11",
"f9",
"f2"
],
"red-team appellate counsel": [
"f10",
"f4",
"f12",
"f2"
],
"brief drafter": [
"f10",
"f3",
"f6",
"f1"
],
"opposing-counsel brief reviewer": [
"f10",
"f5"... | I need you to set up a 5-agent pipeline for civil appellate litigation support. A matter arrives as a trial record excerpt and issue statement in `record.md`, and the pipeline turns it into a vetted appellate theory and then a polished, record-supported appellate brief.
The five agents:
- **docket coordinator** orche... | I need you to set up a 5-agent pipeline for civil appellate litigation support. A matter arrives as a trial record excerpt and issue statement in `record.md`, and the pipeline turns it into a vetted appellate theory and then a polished, record-supported appellate brief.
The five agents:
- **docket coordinator** orche... |
pg_056__seed_1 | pg_056 | 1 | [
"care-coordination lead",
"diagnostic consultant",
"diagnostic safety reviewer",
"care-plan author",
"clinical chart auditor"
] | [
{
"id": "f1",
"text": "Keep `CARE_PLAN.md` lean. Its body — everything except the `<review>` block — must stay under 3000 words (~20KB), about a four-page note; record only the final state and move iteration-by-iteration history into a separate `CHANGELOG.md`. Bold text must not exceed 5% of the file: if ev... | f7 | {
"care-coordination lead": [
"f2"
],
"diagnostic consultant": [
"f3",
"f10",
"f9",
"f5"
],
"diagnostic safety reviewer": [
"f3",
"f4",
"f8",
"f5"
],
"care-plan author": [
"f3",
"f12",
"f6",
"f1"
],
"clinical chart auditor": [
"f3",
"f11",
... | I need you to set up a 5-agent hospital second-opinion workflow that turns a patient presentation into a defensible differential diagnosis and workup strategy, then into a reviewed care-plan note. A case arrives as a clinical presentation in `patient.md`, and the workflow produces a vetted diagnostic plan and then a pr... | I need you to set up a 5-agent hospital second-opinion workflow that turns a patient presentation into a defensible differential diagnosis and workup strategy, then into a reviewed care-plan note. A case arrives as a clinical presentation in `patient.md`, and the workflow produces a vetted diagnostic plan and then a pr... |
pg_056__seed_42 | pg_056 | 42 | [
"care-coordination lead",
"diagnostic consultant",
"diagnostic safety reviewer",
"care-plan author",
"clinical chart auditor"
] | [
{
"id": "f1",
"text": "Keep `CARE_PLAN.md` lean. Its body — everything except the `<review>` block — must stay under 3000 words (~20KB), about a four-page note; record only the final state and move iteration-by-iteration history into a separate `CHANGELOG.md`. Bold text must not exceed 5% of the file: if ev... | f8 | {
"care-coordination lead": [
"f7"
],
"diagnostic consultant": [
"f10",
"f11",
"f9",
"f2"
],
"diagnostic safety reviewer": [
"f10",
"f4",
"f12",
"f2"
],
"care-plan author": [
"f10",
"f3",
"f6",
"f1"
],
"clinical chart auditor": [
"f10",
"f5",... | I need you to set up a 5-agent hospital second-opinion workflow that turns a patient presentation into a defensible differential diagnosis and workup strategy, then into a reviewed care-plan note. A case arrives as a clinical presentation in `patient.md`, and the workflow produces a vetted diagnostic plan and then a pr... | I need you to set up a 5-agent hospital second-opinion workflow that turns a patient presentation into a defensible differential diagnosis and workup strategy, then into a reviewed care-plan note. A case arrives as a clinical presentation in `patient.md`, and the workflow produces a vetted diagnostic plan and then a pr... |
pg_057__seed_1 | pg_057 | 1 | [
"assigning editor",
"investigative reporter",
"skeptical investigations editor",
"story writer",
"fact-checking editor"
] | [
{
"id": "f1",
"text": "Keep `STORY.md` lean. Its body — everything except the `<review>` block — must stay under 3000 words (~20KB), about a four-page note; record only the final state and move iteration-by-iteration reporting history into a separate `CHANGELOG.md`. Bold text must not exceed 5% of the file:... | f7 | {
"assigning editor": [
"f2"
],
"investigative reporter": [
"f3",
"f10",
"f9",
"f5"
],
"skeptical investigations editor": [
"f3",
"f4",
"f8",
"f5"
],
"story writer": [
"f3",
"f12",
"f6",
"f1"
],
"fact-checking editor": [
"f3",
"f11",
"f8"... | I need you to set up a 5-agent investigative journalism publication workflow. A source packet arrives as a collection of tips, documents, interview notes, data extracts, and background material in `source_packet.md`, and the workflow turns it into a vetted reporting theory and then a publishable, fact-checked story.
T... | I need you to set up a 5-agent investigative journalism publication workflow. A source packet arrives as a collection of tips, documents, interview notes, data extracts, and background material in `source_packet.md`, and the workflow turns it into a vetted reporting theory and then a publishable, fact-checked story.
T... |
pg_057__seed_42 | pg_057 | 42 | [
"assigning editor",
"investigative reporter",
"skeptical investigations editor",
"story writer",
"fact-checking editor"
] | [
{
"id": "f1",
"text": "Keep `STORY.md` lean. Its body — everything except the `<review>` block — must stay under 3000 words (~20KB), about a four-page note; record only the final state and move iteration-by-iteration reporting history into a separate `CHANGELOG.md`. Bold text must not exceed 5% of the file:... | f8 | {
"assigning editor": [
"f7"
],
"investigative reporter": [
"f10",
"f11",
"f9",
"f2"
],
"skeptical investigations editor": [
"f10",
"f4",
"f12",
"f2"
],
"story writer": [
"f10",
"f3",
"f6",
"f1"
],
"fact-checking editor": [
"f10",
"f5",
"... | I need you to set up a 5-agent investigative journalism publication workflow. A source packet arrives as a collection of tips, documents, interview notes, data extracts, and background material in `source_packet.md`, and the workflow turns it into a vetted reporting theory and then a publishable, fact-checked story.
T... | I need you to set up a 5-agent investigative journalism publication workflow. A source packet arrives as a collection of tips, documents, interview notes, data extracts, and background material in `source_packet.md`, and the workflow turns it into a vetted reporting theory and then a publishable, fact-checked story.
T... |
pg_058__seed_1 | pg_058 | 1 | [
"collections coordinator",
"provenance researcher",
"forgery-risk reviewer",
"acquisition dossier writer",
"due-diligence auditor"
] | [
{
"id": "f1",
"text": "Keep `DOSSIER.md` lean. Its body — everything except the `<review>` block — must stay under 3000 words (~20KB), about a four-page committee note; record only the final state and move iteration-by-iteration history into a separate `CHANGELOG.md`. Bold text must not exceed 5% of the fil... | f7 | {
"collections coordinator": [
"f2"
],
"provenance researcher": [
"f3",
"f10",
"f9",
"f5"
],
"forgery-risk reviewer": [
"f3",
"f4",
"f8",
"f5"
],
"acquisition dossier writer": [
"f3",
"f12",
"f6",
"f1"
],
"due-diligence auditor": [
"f3",
"f11... | I need you to set up a 5-agent pipeline for museum acquisition due diligence. An offered object's file arrives as a descriptive submission in `object_file.md`, and the pipeline turns it into a vetted authentication and provenance assessment and then a complete acquisition dossier ready for committee review.
The five a... | I need you to set up a 5-agent pipeline for museum acquisition due diligence. An offered object's file arrives as a descriptive submission in `object_file.md`, and the pipeline turns it into a vetted authentication and provenance assessment and then a complete acquisition dossier ready for committee review.
The five a... |
pg_058__seed_42 | pg_058 | 42 | [
"collections coordinator",
"provenance researcher",
"forgery-risk reviewer",
"acquisition dossier writer",
"due-diligence auditor"
] | [
{
"id": "f1",
"text": "Keep `DOSSIER.md` lean. Its body — everything except the `<review>` block — must stay under 3000 words (~20KB), about a four-page committee note; record only the final state and move iteration-by-iteration history into a separate `CHANGELOG.md`. Bold text must not exceed 5% of the fil... | f8 | {
"collections coordinator": [
"f7"
],
"provenance researcher": [
"f10",
"f11",
"f9",
"f2"
],
"forgery-risk reviewer": [
"f10",
"f4",
"f12",
"f2"
],
"acquisition dossier writer": [
"f10",
"f3",
"f6",
"f1"
],
"due-diligence auditor": [
"f10",
... | I need you to set up a 5-agent pipeline for museum acquisition due diligence. An offered object's file arrives as a descriptive submission in `object_file.md`, and the pipeline turns it into a vetted authentication and provenance assessment and then a complete acquisition dossier ready for committee review.
The five a... | I need you to set up a 5-agent pipeline for museum acquisition due diligence. An offered object's file arrives as a descriptive submission in `object_file.md`, and the pipeline turns it into a vetted authentication and provenance assessment and then a complete acquisition dossier ready for committee review.
The five a... |
pg_059__seed_1 | pg_059 | 1 | [
"permitting coordinator",
"impact assessment lead",
"regulatory challenge reviewer",
"EIA report author",
"permit-compliance auditor"
] | [
{
"id": "f1",
"text": "Keep `EIA_REPORT.md` lean. Its body — everything except the `<review>` block — must stay under 3000 words (~20KB), about a four-page permit memorandum; record only the final state and move iteration-by-iteration history into a separate `CHANGELOG.md`. Bold text must not exceed 5% of t... | f7 | {
"permitting coordinator": [
"f2"
],
"impact assessment lead": [
"f3",
"f10",
"f9",
"f5"
],
"regulatory challenge reviewer": [
"f3",
"f4",
"f8",
"f5"
],
"EIA report author": [
"f3",
"f12",
"f6",
"f1"
],
"permit-compliance auditor": [
"f3",
"... | I need you to set up a 5-agent pipeline for an environmental impact assessment for a proposed construction project. Project materials arrive as a natural-language package in `project.md`, and the pipeline turns them into a vetted impact-analysis approach and then a permit-ready EIA report that can withstand regulator-s... | I need you to set up a 5-agent pipeline for an environmental impact assessment for a proposed construction project. Project materials arrive as a natural-language package in `project.md`, and the pipeline turns them into a vetted impact-analysis approach and then a permit-ready EIA report that can withstand regulator-s... |
pg_059__seed_42 | pg_059 | 42 | [
"permitting coordinator",
"impact assessment lead",
"regulatory challenge reviewer",
"EIA report author",
"permit-compliance auditor"
] | [
{
"id": "f1",
"text": "Keep `EIA_REPORT.md` lean. Its body — everything except the `<review>` block — must stay under 3000 words (~20KB), about a four-page permit memorandum; record only the final state and move iteration-by-iteration history into a separate `CHANGELOG.md`. Bold text must not exceed 5% of t... | f8 | {
"permitting coordinator": [
"f7"
],
"impact assessment lead": [
"f10",
"f11",
"f9",
"f2"
],
"regulatory challenge reviewer": [
"f10",
"f4",
"f12",
"f2"
],
"EIA report author": [
"f10",
"f3",
"f6",
"f1"
],
"permit-compliance auditor": [
"f10",
... | I need you to set up a 5-agent pipeline for an environmental impact assessment for a proposed construction project. Project materials arrive as a natural-language package in `project.md`, and the pipeline turns them into a vetted impact-analysis approach and then a permit-ready EIA report that can withstand regulator-s... | I need you to set up a 5-agent pipeline for an environmental impact assessment for a proposed construction project. Project materials arrive as a natural-language package in `project.md`, and the pipeline turns them into a vetted impact-analysis approach and then a permit-ready EIA report that can withstand regulator-s... |
pg_060__seed_1 | pg_060 | 1 | [
"litigation-manager",
"lead-strategist",
"argument-developer",
"moot-court-judge",
"motion-drafter",
"opposing-counsel-proxy"
] | [
{
"id": "f1",
"text": "Keep your assigned plan's `SOLUTION.md` lean. Its body — everything except the `<review>` block — must stay under 3000 words (~20KB), about a twelve-page brief; record only the final draft and move iteration-by-iteration history into the plan's `CHANGELOG.md`. Bold text must not excee... | f7 | {
"litigation-manager": [
"f2"
],
"lead-strategist": [
"f3",
"f9",
"f5"
],
"argument-developer": [
"f3",
"f5",
"f10"
],
"moot-court-judge": [
"f3",
"f5",
"f4",
"f8"
],
"motion-drafter": [
"f3",
"f6",
"f1",
"f12"
],
"opposing-counsel-proxy... | I need you to set up a 6-agent pipeline that drafts a comprehensive legal defense strategy and finalizes the exact court motions to survive summary judgment in a high-stakes corporate litigation case. A case arrives as a natural-language statement in `problem.md`, and the pipeline turns it into a vetted plan and then f... | I need you to set up a 6-agent pipeline that drafts a comprehensive legal defense strategy and finalizes the exact court motions to survive summary judgment in a high-stakes corporate litigation case. A case arrives as a natural-language statement in `problem.md`, and the pipeline turns it into a vetted plan and then f... |
pg_060__seed_42 | pg_060 | 42 | [
"litigation-manager",
"lead-strategist",
"argument-developer",
"moot-court-judge",
"motion-drafter",
"opposing-counsel-proxy"
] | [
{
"id": "f1",
"text": "Keep your assigned plan's `SOLUTION.md` lean. Its body — everything except the `<review>` block — must stay under 3000 words (~20KB), about a twelve-page brief; record only the final draft and move iteration-by-iteration history into the plan's `CHANGELOG.md`. Bold text must not excee... | f8 | {
"litigation-manager": [
"f7"
],
"lead-strategist": [
"f10",
"f9",
"f2"
],
"argument-developer": [
"f10",
"f2",
"f11"
],
"moot-court-judge": [
"f10",
"f2",
"f4",
"f12"
],
"motion-drafter": [
"f10",
"f6",
"f1",
"f3"
],
"opposing-counsel-p... | I need you to set up a 6-agent pipeline that drafts a comprehensive legal defense strategy and finalizes the exact court motions to survive summary judgment in a high-stakes corporate litigation case. A case arrives as a natural-language statement in `problem.md`, and the pipeline turns it into a vetted plan and then f... | I need you to set up a 6-agent pipeline that drafts a comprehensive legal defense strategy and finalizes the exact court motions to survive summary judgment in a high-stakes corporate litigation case. A case arrives as a natural-language statement in `problem.md`, and the pipeline turns it into a vetted plan and then f... |
pg_061__seed_1 | pg_061 | 1 | [
"research-director",
"pathway-architect",
"molecule-optimizer",
"binding-skeptic",
"protocol-author",
"toxicology-auditor"
] | [
{
"id": "f1",
"text": "Keep your assigned plan's `SOLUTION.md` lean. Its body — everything except the `<review>` block — must stay under 3000 words (~20KB), about a four-page note; record only the final state and move iteration-by-iteration history into the plan's `CHANGELOG.md`. Bold text must not exceed 5... | f7 | {
"research-director": [
"f2"
],
"pathway-architect": [
"f3",
"f9",
"f5"
],
"molecule-optimizer": [
"f3",
"f5",
"f10"
],
"binding-skeptic": [
"f3",
"f5",
"f4",
"f8"
],
"protocol-author": [
"f3",
"f6",
"f1",
"f12"
],
"toxicology-auditor": ... | I need you to set up a 6-agent pipeline that designs a novel biochemical therapeutic targeting a specific evaluationular receptor and details the rigorous laboratory synthesis protocol to safely produce it. A therapeutic target arrives as a natural-language statement in `problem.md`, and the pipeline turns it into a ve... | I need you to set up a 6-agent pipeline that designs a novel biochemical therapeutic targeting a specific evaluationular receptor and details the rigorous laboratory synthesis protocol to safely produce it. A therapeutic target arrives as a natural-language statement in `problem.md`, and the pipeline turns it into a ve... |
pg_061__seed_42 | pg_061 | 42 | [
"research-director",
"pathway-architect",
"molecule-optimizer",
"binding-skeptic",
"protocol-author",
"toxicology-auditor"
] | [
{
"id": "f1",
"text": "Keep your assigned plan's `SOLUTION.md` lean. Its body — everything except the `<review>` block — must stay under 3000 words (~20KB), about a four-page note; record only the final state and move iteration-by-iteration history into the plan's `CHANGELOG.md`. Bold text must not exceed 5... | f8 | {
"research-director": [
"f7"
],
"pathway-architect": [
"f10",
"f9",
"f2"
],
"molecule-optimizer": [
"f10",
"f2",
"f11"
],
"binding-skeptic": [
"f10",
"f2",
"f4",
"f12"
],
"protocol-author": [
"f10",
"f6",
"f1",
"f3"
],
"toxicology-audito... | I need you to set up a 6-agent pipeline that designs a novel biochemical therapeutic targeting a specific evaluationular receptor and details the rigorous laboratory synthesis protocol to safely produce it. A therapeutic target arrives as a natural-language statement in `problem.md`, and the pipeline turns it into a ve... | I need you to set up a 6-agent pipeline that designs a novel biochemical therapeutic targeting a specific evaluationular receptor and details the rigorous laboratory synthesis protocol to safely produce it. A therapeutic target arrives as a natural-language statement in `problem.md`, and the pipeline turns it into a ve... |
pg_062__seed_1 | pg_062 | 1 | [
"operations-director",
"tactical-theorist",
"scenario-adapter",
"red-team-analyst",
"field-schedule-writer",
"counter-intel-auditor"
] | [
{
"id": "f1",
"text": "Keep your assigned plan's `SOLUTION.md` lean. Its body — everything except the `<review>` block — must stay under 3000 words (~20KB), about a four-page operations memo; record only the final logistical state and move iteration-by-iteration planning history into the plan's `CHANGELOG.m... | f7 | {
"operations-director": [
"f2"
],
"tactical-theorist": [
"f3",
"f9",
"f5"
],
"scenario-adapter": [
"f3",
"f5",
"f10"
],
"red-team-analyst": [
"f3",
"f5",
"f4",
"f8"
],
"field-schedule-writer": [
"f3",
"f6",
"f1",
"f12"
],
"counter-intel-... | I need you to set up a 6-agent pipeline that plans high-risk covert intelligence extraction operations and generates minute-by-minute logistical field execution orders. A problem arrives as a natural-language mission brief in `problem.md`, and the pipeline turns it into a vetted operation plan and then working, verifie... | I need you to set up a 6-agent pipeline that plans high-risk covert intelligence extraction operations and generates minute-by-minute logistical field execution orders. A problem arrives as a natural-language mission brief in `problem.md`, and the pipeline turns it into a vetted operation plan and then working, verifie... |
pg_062__seed_42 | pg_062 | 42 | [
"operations-director",
"tactical-theorist",
"scenario-adapter",
"red-team-analyst",
"field-schedule-writer",
"counter-intel-auditor"
] | [
{
"id": "f1",
"text": "Keep your assigned plan's `SOLUTION.md` lean. Its body — everything except the `<review>` block — must stay under 3000 words (~20KB), about a four-page operations memo; record only the final logistical state and move iteration-by-iteration planning history into the plan's `CHANGELOG.m... | f8 | {
"operations-director": [
"f7"
],
"tactical-theorist": [
"f10",
"f9",
"f2"
],
"scenario-adapter": [
"f10",
"f2",
"f11"
],
"red-team-analyst": [
"f10",
"f2",
"f4",
"f12"
],
"field-schedule-writer": [
"f10",
"f6",
"f1",
"f3"
],
"counter-in... | I need you to set up a 6-agent pipeline that plans high-risk covert intelligence extraction operations and generates minute-by-minute logistical field execution orders. A problem arrives as a natural-language mission brief in `problem.md`, and the pipeline turns it into a vetted operation plan and then working, verifie... | I need you to set up a 6-agent pipeline that plans high-risk covert intelligence extraction operations and generates minute-by-minute logistical field execution orders. A problem arrives as a natural-language mission brief in `problem.md`, and the pipeline turns it into a vetted operation plan and then working, verifie... |
pg_063__seed_1 | pg_063 | 1 | [
"policy-coordinator",
"macro-theorist",
"intervention-tuner",
"economic-skeptic",
"legislation-drafter",
"compliance-stress-tester"
] | [
{
"id": "f1",
"text": "Keep your assigned plan's `SOLUTION.md` lean. Its body — everything except the `<review>` block — must stay under 3000 words (~20KB), about a four-page legal brief; record only the final drafted text and move iteration-by-iteration legislative history into the plan's `CHANGELOG.md`. B... | f7 | {
"policy-coordinator": [
"f2"
],
"macro-theorist": [
"f3",
"f9",
"f5"
],
"intervention-tuner": [
"f3",
"f5",
"f10"
],
"economic-skeptic": [
"f3",
"f5",
"f4",
"f8"
],
"legislation-drafter": [
"f3",
"f6",
"f1",
"f12"
],
"compliance-stress-... | I need you to set up a 6-agent pipeline that engineers a macroeconomic intervention to stabilize a collapsing national currency and drafts the specific legislative mandates and banking regulations to enforce it. A crisis brief arrives as a natural-language statement in `problem.md`, and the pipeline turns it into a vet... | I need you to set up a 6-agent pipeline that engineers a macroeconomic intervention to stabilize a collapsing national currency and drafts the specific legislative mandates and banking regulations to enforce it. A crisis brief arrives as a natural-language statement in `problem.md`, and the pipeline turns it into a vet... |
pg_063__seed_42 | pg_063 | 42 | [
"policy-coordinator",
"macro-theorist",
"intervention-tuner",
"economic-skeptic",
"legislation-drafter",
"compliance-stress-tester"
] | [
{
"id": "f1",
"text": "Keep your assigned plan's `SOLUTION.md` lean. Its body — everything except the `<review>` block — must stay under 3000 words (~20KB), about a four-page legal brief; record only the final drafted text and move iteration-by-iteration legislative history into the plan's `CHANGELOG.md`. B... | f8 | {
"policy-coordinator": [
"f7"
],
"macro-theorist": [
"f10",
"f9",
"f2"
],
"intervention-tuner": [
"f10",
"f2",
"f11"
],
"economic-skeptic": [
"f10",
"f2",
"f4",
"f12"
],
"legislation-drafter": [
"f10",
"f6",
"f1",
"f3"
],
"compliance-str... | I need you to set up a 6-agent pipeline that engineers a macroeconomic intervention to stabilize a collapsing national currency and drafts the specific legislative mandates and banking regulations to enforce it. A crisis brief arrives as a natural-language statement in `problem.md`, and the pipeline turns it into a vet... | I need you to set up a 6-agent pipeline that engineers a macroeconomic intervention to stabilize a collapsing national currency and drafts the specific legislative mandates and banking regulations to enforce it. A crisis brief arrives as a natural-language statement in `problem.md`, and the pipeline turns it into a vet... |
pg_064__seed_1 | pg_064 | 1 | [
"mission-commander",
"astrodynamics-architect",
"maneuver-adjuster",
"orbital-mechanics-reviewer",
"burn-sequence-compiler",
"physics-simulator-proxy"
] | [
{
"id": "f1",
"text": "Keep your assigned plan's `SOLUTION.md` lean. Its body — everything except the `<review>` block — must stay under 3000 words (~20KB), about a four-page note; record only the final sequence and move iteration-by-iteration telemetry history into the plan's `CHANGELOG.md`. Bold text must... | f7 | {
"mission-commander": [
"f2"
],
"astrodynamics-architect": [
"f3",
"f9",
"f5"
],
"maneuver-adjuster": [
"f3",
"f5",
"f10"
],
"orbital-mechanics-reviewer": [
"f3",
"f5",
"f4",
"f8"
],
"burn-sequence-compiler": [
"f3",
"f6",
"f1",
"f12"
],
... | I need you to set up a 6-agent pipeline that designs a deep-space probe trajectory and compiles the exact thruster burn sequences, timings, and telemetry expectations for the flight computer. A mission profile arrives as a natural-language statement in `problem.md`, and the pipeline turns it into a vetted trajectory pl... | I need you to set up a 6-agent pipeline that designs a deep-space probe trajectory and compiles the exact thruster burn sequences, timings, and telemetry expectations for the flight computer. A mission profile arrives as a natural-language statement in `problem.md`, and the pipeline turns it into a vetted trajectory pl... |
pg_064__seed_42 | pg_064 | 42 | [
"mission-commander",
"astrodynamics-architect",
"maneuver-adjuster",
"orbital-mechanics-reviewer",
"burn-sequence-compiler",
"physics-simulator-proxy"
] | [
{
"id": "f1",
"text": "Keep your assigned plan's `SOLUTION.md` lean. Its body — everything except the `<review>` block — must stay under 3000 words (~20KB), about a four-page note; record only the final sequence and move iteration-by-iteration telemetry history into the plan's `CHANGELOG.md`. Bold text must... | f8 | {
"mission-commander": [
"f7"
],
"astrodynamics-architect": [
"f10",
"f9",
"f2"
],
"maneuver-adjuster": [
"f10",
"f2",
"f11"
],
"orbital-mechanics-reviewer": [
"f10",
"f2",
"f4",
"f12"
],
"burn-sequence-compiler": [
"f10",
"f6",
"f1",
"f3"
... | I need you to set up a 6-agent pipeline that designs a deep-space probe trajectory and compiles the exact thruster burn sequences, timings, and telemetry expectations for the flight computer. A mission profile arrives as a natural-language statement in `problem.md`, and the pipeline turns it into a vetted trajectory pl... | I need you to set up a 6-agent pipeline that designs a deep-space probe trajectory and compiles the exact thruster burn sequences, timings, and telemetry expectations for the flight computer. A mission profile arrives as a natural-language statement in `problem.md`, and the pipeline turns it into a vetted trajectory pl... |
pg_065__seed_1 | pg_065 | 1 | [
"case coordinator",
"litigation strategist",
"theory reviser",
"opposing-counsel analyst",
"brief drafter",
"brief challenge reviewer"
] | [
{
"id": "f1",
"text": "Keep your assigned theory's `BRIEF.md` lean. Its body — everything except the `<review>` block — must stay under 3000 words (~20KB), about a four-page note; record only the final state and move iteration-by-iteration history into the theory's `CHANGELOG.md`. Bold text must not exceed ... | f7 | {
"case coordinator": [
"f2"
],
"litigation strategist": [
"f3",
"f9",
"f5"
],
"theory reviser": [
"f3",
"f5",
"f10"
],
"opposing-counsel analyst": [
"f3",
"f5",
"f4",
"f8"
],
"brief drafter": [
"f3",
"f6",
"f1",
"f12"
],
"brief challenge... | I need you to set up a 6-agent pipeline for civil litigation preparation. A case arrives as a factual and procedural summary in `casefile.md`, and the pipeline turns it into vetted legal theories and then a working motion or brief that has been adversarially reviewed for weaknesses.
The six agents:
- **case coordinat... | I need you to set up a 6-agent pipeline for civil litigation preparation. A case arrives as a factual and procedural summary in `casefile.md`, and the pipeline turns it into vetted legal theories and then a working motion or brief that has been adversarially reviewed for weaknesses.
The six agents:
- **case coordinat... |
pg_065__seed_42 | pg_065 | 42 | [
"case coordinator",
"litigation strategist",
"theory reviser",
"opposing-counsel analyst",
"brief drafter",
"brief challenge reviewer"
] | [
{
"id": "f1",
"text": "Keep your assigned theory's `BRIEF.md` lean. Its body — everything except the `<review>` block — must stay under 3000 words (~20KB), about a four-page note; record only the final state and move iteration-by-iteration history into the theory's `CHANGELOG.md`. Bold text must not exceed ... | f8 | {
"case coordinator": [
"f7"
],
"litigation strategist": [
"f10",
"f9",
"f2"
],
"theory reviser": [
"f10",
"f2",
"f11"
],
"opposing-counsel analyst": [
"f10",
"f2",
"f4",
"f12"
],
"brief drafter": [
"f10",
"f6",
"f1",
"f3"
],
"brief chall... | I need you to set up a 6-agent pipeline for civil litigation preparation. A case arrives as a factual and procedural summary in `casefile.md`, and the pipeline turns it into vetted legal theories and then a working motion or brief that has been adversarially reviewed for weaknesses.
The six agents:
- **case coordinat... | I need you to set up a 6-agent pipeline for civil litigation preparation. A case arrives as a factual and procedural summary in `casefile.md`, and the pipeline turns it into vetted legal theories and then a working motion or brief that has been adversarially reviewed for weaknesses.
The six agents:
- **case coordinat... |
pg_066__seed_1 | pg_066 | 1 | [
"collections coordinator",
"provenance hypothesis curator",
"attribution reviser",
"forgery-risk examiner",
"catalog report author",
"authentication audit reviewer"
] | [
{
"id": "f1",
"text": "Keep your assigned hypothesis's `ATTRIBUTION_REPORT.md` lean. Its body — everything except the `<review>` block — must stay under 3000 words (~20KB), about a four-page catalog note; record only the final state and move iteration-by-iteration history into the hypothesis's `CHANGELOG.md... | f7 | {
"collections coordinator": [
"f2"
],
"provenance hypothesis curator": [
"f3",
"f9",
"f5"
],
"attribution reviser": [
"f3",
"f5",
"f10"
],
"forgery-risk examiner": [
"f3",
"f5",
"f4",
"f8"
],
"catalog report author": [
"f3",
"f6",
"f1",
"f12... | I need you to set up a 6-agent pipeline that evaluates a museum artifact's authenticity and attribution. An object arrives with a natural-language dossier in `artifact_dossier.md`, and the pipeline turns it into competing provenance hypotheses, a vetted hypothesis, and then a defensible attribution report that has been... | I need you to set up a 6-agent pipeline that evaluates a museum artifact's authenticity and attribution. An object arrives with a natural-language dossier in `artifact_dossier.md`, and the pipeline turns it into competing provenance hypotheses, a vetted hypothesis, and then a defensible attribution report that has been... |
pg_066__seed_42 | pg_066 | 42 | [
"collections coordinator",
"provenance hypothesis curator",
"attribution reviser",
"forgery-risk examiner",
"catalog report author",
"authentication audit reviewer"
] | [
{
"id": "f1",
"text": "Keep your assigned hypothesis's `ATTRIBUTION_REPORT.md` lean. Its body — everything except the `<review>` block — must stay under 3000 words (~20KB), about a four-page catalog note; record only the final state and move iteration-by-iteration history into the hypothesis's `CHANGELOG.md... | f8 | {
"collections coordinator": [
"f7"
],
"provenance hypothesis curator": [
"f10",
"f9",
"f2"
],
"attribution reviser": [
"f10",
"f2",
"f11"
],
"forgery-risk examiner": [
"f10",
"f2",
"f4",
"f12"
],
"catalog report author": [
"f10",
"f6",
"f1",
... | I need you to set up a 6-agent pipeline that evaluates a museum artifact's authenticity and attribution. An object arrives with a natural-language dossier in `artifact_dossier.md`, and the pipeline turns it into competing provenance hypotheses, a vetted hypothesis, and then a defensible attribution report that has been... | I need you to set up a 6-agent pipeline that evaluates a museum artifact's authenticity and attribution. An object arrives with a natural-language dossier in `artifact_dossier.md`, and the pipeline turns it into competing provenance hypotheses, a vetted hypothesis, and then a defensible attribution report that has been... |
pg_067__seed_1 | pg_067 | 1 | [
"case-flow coordinator",
"diagnostic pathway proposer",
"care-plan reviser",
"patient-safety challenger",
"clinical recommendation writer",
"medical audit reviewer"
] | [
{
"id": "f1",
"text": "Keep your assigned pathway's `RECOMMENDATION.md` lean. Its body — everything except the `<review>` block — must stay under 3000 words (~20KB), about a four-page note; record only the final state and move iteration-by-iteration history into the pathway's `CHANGELOG.md`. Bold text must ... | f7 | {
"case-flow coordinator": [
"f2"
],
"diagnostic pathway proposer": [
"f3",
"f9",
"f5"
],
"care-plan reviser": [
"f3",
"f5",
"f10"
],
"patient-safety challenger": [
"f3",
"f5",
"f4",
"f8"
],
"clinical recommendation writer": [
"f3",
"f6",
"f1",
... | I need you to set up a 6-agent pipeline that reviews clinical cases. A case arrives as a structured clinical narrative in `case.md`, and the pipeline turns it into several competing diagnostic and treatment pathways, a safety-vetted care plan, and then a patient-specific care recommendation audited for medical errors, ... | I need you to set up a 6-agent pipeline that reviews clinical cases. A case arrives as a structured clinical narrative in `case.md`, and the pipeline turns it into several competing diagnostic and treatment pathways, a safety-vetted care plan, and then a patient-specific care recommendation audited for medical errors, ... |
pg_067__seed_42 | pg_067 | 42 | [
"case-flow coordinator",
"diagnostic pathway proposer",
"care-plan reviser",
"patient-safety challenger",
"clinical recommendation writer",
"medical audit reviewer"
] | [
{
"id": "f1",
"text": "Keep your assigned pathway's `RECOMMENDATION.md` lean. Its body — everything except the `<review>` block — must stay under 3000 words (~20KB), about a four-page note; record only the final state and move iteration-by-iteration history into the pathway's `CHANGELOG.md`. Bold text must ... | f8 | {
"case-flow coordinator": [
"f7"
],
"diagnostic pathway proposer": [
"f10",
"f9",
"f2"
],
"care-plan reviser": [
"f10",
"f2",
"f11"
],
"patient-safety challenger": [
"f10",
"f2",
"f4",
"f12"
],
"clinical recommendation writer": [
"f10",
"f6",
"f... | I need you to set up a 6-agent pipeline that reviews clinical cases. A case arrives as a structured clinical narrative in `case.md`, and the pipeline turns it into several competing diagnostic and treatment pathways, a safety-vetted care plan, and then a patient-specific care recommendation audited for medical errors, ... | I need you to set up a 6-agent pipeline that reviews clinical cases. A case arrives as a structured clinical narrative in `case.md`, and the pipeline turns it into several competing diagnostic and treatment pathways, a safety-vetted care plan, and then a patient-specific care recommendation audited for medical errors, ... |
pg_068__seed_1 | pg_068 | 1 | [
"project coordinator",
"structural scheme designer",
"design revision engineer",
"code-compliance examiner",
"calculation package engineer",
"independent structural reviewer"
] | [
{
"id": "f1",
"text": "Keep your assigned scheme's `CALC_PACKAGE.md` lean. Its body — everything except the `<review>` block — must stay under 3000 words (~20KB), about a four-page calculation narrative; record only the final state and move iteration-by-iteration history into the scheme's `CHANGELOG.md`. Bo... | f7 | {
"project coordinator": [
"f2"
],
"structural scheme designer": [
"f3",
"f9",
"f5"
],
"design revision engineer": [
"f3",
"f5",
"f10"
],
"code-compliance examiner": [
"f3",
"f5",
"f4",
"f8"
],
"calculation package engineer": [
"f3",
"f6",
"f1",
... | I need you to set up a 6-agent pipeline for structural building design compliance. A project arrives as a natural-language design brief in `project_brief.md`, and the pipeline turns it into vetted alternative structural schemes, then into a working calculation package that is independently reviewed for building-code vi... | I need you to set up a 6-agent pipeline for structural building design compliance. A project arrives as a natural-language design brief in `project_brief.md`, and the pipeline turns it into vetted alternative structural schemes, then into a working calculation package that is independently reviewed for building-code vi... |
pg_068__seed_42 | pg_068 | 42 | [
"project coordinator",
"structural scheme designer",
"design revision engineer",
"code-compliance examiner",
"calculation package engineer",
"independent structural reviewer"
] | [
{
"id": "f1",
"text": "Keep your assigned scheme's `CALC_PACKAGE.md` lean. Its body — everything except the `<review>` block — must stay under 3000 words (~20KB), about a four-page calculation narrative; record only the final state and move iteration-by-iteration history into the scheme's `CHANGELOG.md`. Bo... | f8 | {
"project coordinator": [
"f7"
],
"structural scheme designer": [
"f10",
"f9",
"f2"
],
"design revision engineer": [
"f10",
"f2",
"f11"
],
"code-compliance examiner": [
"f10",
"f2",
"f4",
"f12"
],
"calculation package engineer": [
"f10",
"f6",
"... | I need you to set up a 6-agent pipeline for structural building design compliance. A project arrives as a natural-language design brief in `project_brief.md`, and the pipeline turns it into vetted alternative structural schemes, then into a working calculation package that is independently reviewed for building-code vi... | I need you to set up a 6-agent pipeline for structural building design compliance. A project arrives as a natural-language design brief in `project_brief.md`, and the pipeline turns it into vetted alternative structural schemes, then into a working calculation package that is independently reviewed for building-code vi... |
pg_069__seed_1 | pg_069 | 1 | [
"editorial coordinator",
"investigation angle developer",
"reporting plan reviser",
"skeptical editor",
"story writer",
"fact-check editor"
] | [
{
"id": "f1",
"text": "Keep your assigned angle's `ARTICLE.md` lean. Its body — everything except the `<review>` block — must stay under 3000 words (~20KB), about a four-page draft and sourcing note; record iteration-by-iteration reporting history in the angle's `CHANGELOG.md`. Bold text must not exceed 5% ... | f7 | {
"editorial coordinator": [
"f2"
],
"investigation angle developer": [
"f3",
"f9",
"f5"
],
"reporting plan reviser": [
"f3",
"f5",
"f10"
],
"skeptical editor": [
"f3",
"f5",
"f4",
"f8"
],
"story writer": [
"f3",
"f6",
"f1",
"f12"
],
"fac... | I need you to set up a 6-agent pipeline for investigative journalism fact-checking. A lead, tip, or assignment arrives as a written brief in `assignment.md`, and the pipeline turns it into competing vetted investigation angles, then into a publishable article draft that has been adversarially fact-checked before public... | I need you to set up a 6-agent pipeline for investigative journalism fact-checking. A lead, tip, or assignment arrives as a written brief in `assignment.md`, and the pipeline turns it into competing vetted investigation angles, then into a publishable article draft that has been adversarially fact-checked before public... |
pg_069__seed_42 | pg_069 | 42 | [
"editorial coordinator",
"investigation angle developer",
"reporting plan reviser",
"skeptical editor",
"story writer",
"fact-check editor"
] | [
{
"id": "f1",
"text": "Keep your assigned angle's `ARTICLE.md` lean. Its body — everything except the `<review>` block — must stay under 3000 words (~20KB), about a four-page draft and sourcing note; record iteration-by-iteration reporting history in the angle's `CHANGELOG.md`. Bold text must not exceed 5% ... | f8 | {
"editorial coordinator": [
"f7"
],
"investigation angle developer": [
"f10",
"f9",
"f2"
],
"reporting plan reviser": [
"f10",
"f2",
"f11"
],
"skeptical editor": [
"f10",
"f2",
"f4",
"f12"
],
"story writer": [
"f10",
"f6",
"f1",
"f3"
],
... | I need you to set up a 6-agent pipeline for investigative journalism fact-checking. A lead, tip, or assignment arrives as a written brief in `assignment.md`, and the pipeline turns it into competing vetted investigation angles, then into a publishable article draft that has been adversarially fact-checked before public... | I need you to set up a 6-agent pipeline for investigative journalism fact-checking. A lead, tip, or assignment arrives as a written brief in `assignment.md`, and the pipeline turns it into competing vetted investigation angles, then into a publishable article draft that has been adversarially fact-checked before public... |
pg_070__seed_1 | pg_070 | 1 | [
"station_chief",
"lead_analyst",
"field_collector",
"red_teamer"
] | [
{
"id": "f1",
"text": "Review `HIGH.md` adversarially with the standards of a national intelligence directorate.\nYou are not part of the operational team and not their internal QA.\nYour job is to reject drift, weak intelligence, and threat claims larger than what the intercepts show.\n\nCover four layers.... | f11 | {
"station_chief": [
"f2"
],
"lead_analyst": [
"f7",
"f6",
"f8",
"f10",
"f9"
],
"field_collector": [
"f7",
"f5",
"f3",
"f9"
],
"red_teamer": [
"f7",
"f1",
"f4",
"f9"
]
} | I need you to set up a 4-agent counter-terrorism intelligence pipeline.
The 4 agents are a station_chief and 3 intelligence agents (lead_analyst, field_collector, red_teamer).
The lead_analyst models threat networks, interprets raw intercepts, and responds to red-team critiques.
The lead_analyst writes the high-level ... | I need you to set up a 4-agent counter-terrorism intelligence pipeline.
The 4 agents are a station_chief and 3 intelligence agents (lead_analyst, field_collector, red_teamer).
The lead_analyst models threat networks, interprets raw intercepts, and responds to red-team critiques.
The lead_analyst writes the high-level ... |
pg_070__seed_42 | pg_070 | 42 | [
"station_chief",
"lead_analyst",
"field_collector",
"red_teamer"
] | [
{
"id": "f1",
"text": "Score the intelligence assessment from 1 to 10.\nRecord the verdict by overwriting the single `<review score=X>` block at the end of `HIGH.md`.\nAlways keep exactly one `<review>` block.\nReplace it, never append a second.\n\nScore strictly on real analytical merit.\nDo not give credi... | f3 | {
"station_chief": [
"f4"
],
"lead_analyst": [
"f9",
"f2",
"f8",
"f7",
"f10"
],
"field_collector": [
"f9",
"f5",
"f11",
"f10"
],
"red_teamer": [
"f9",
"f6",
"f1",
"f10"
]
} | I need you to set up a 4-agent counter-terrorism intelligence pipeline.
The 4 agents are a station_chief and 3 intelligence agents (lead_analyst, field_collector, red_teamer).
The lead_analyst models threat networks, interprets raw intercepts, and responds to red-team critiques.
The lead_analyst writes the high-level ... | I need you to set up a 4-agent counter-terrorism intelligence pipeline.
The 4 agents are a station_chief and 3 intelligence agents (lead_analyst, field_collector, red_teamer).
The lead_analyst models threat networks, interprets raw intercepts, and responds to red-team critiques.
The lead_analyst writes the high-level ... |
pg_071__seed_1 | pg_071 | 1 | [
"managing_partner",
"deal_strategist",
"forensic_accountant",
"risk_auditor"
] | [
{
"id": "f1",
"text": "Audit `HIGH.md` adversarially with the standards of a strict investment committee.\nYou are not part of the deal team and not their internal cheerleader.\nYour job is to reject thesis drift, weak financial evidence, and valuations larger than what the accounting data supports.\n\nCove... | f11 | {
"managing_partner": [
"f2"
],
"deal_strategist": [
"f7",
"f6",
"f8",
"f10",
"f9"
],
"forensic_accountant": [
"f7",
"f5",
"f3",
"f9"
],
"risk_auditor": [
"f7",
"f1",
"f4",
"f9"
]
} | I need you to set up a 4-agent M&A pipeline.
The 4 agents are a managing_partner and 3 deal agents (deal_strategist, forensic_accountant, risk_auditor).
The deal_strategist pitches buyout theses, interprets financial data, and responds to audits.
The deal_strategist writes the high-level deal thesis `HIGH.md`.
The bod... | I need you to set up a 4-agent M&A pipeline.
The 4 agents are a managing_partner and 3 deal agents (deal_strategist, forensic_accountant, risk_auditor).
The deal_strategist pitches buyout theses, interprets financial data, and responds to audits.
The deal_strategist writes the high-level deal thesis `HIGH.md`.
The bod... |
pg_071__seed_42 | pg_071 | 42 | [
"managing_partner",
"deal_strategist",
"forensic_accountant",
"risk_auditor"
] | [
{
"id": "f1",
"text": "Score the deal thesis from 1 to 10.\nRecord the verdict by overwriting the single `<audit score=X>` block at the end of `HIGH.md`.\nAlways keep exactly one `<audit>` block.\nReplace it, never append a second.\n\nScore strictly on financial merit and risk mitigation.\nDo not give credi... | f3 | {
"managing_partner": [
"f4"
],
"deal_strategist": [
"f9",
"f2",
"f8",
"f7",
"f10"
],
"forensic_accountant": [
"f9",
"f5",
"f11",
"f10"
],
"risk_auditor": [
"f9",
"f6",
"f1",
"f10"
]
} | I need you to set up a 4-agent M&A pipeline.
The 4 agents are a managing_partner and 3 deal agents (deal_strategist, forensic_accountant, risk_auditor).
The deal_strategist pitches buyout theses, interprets financial data, and responds to audits.
The deal_strategist writes the high-level deal thesis `HIGH.md`.
The bod... | I need you to set up a 4-agent M&A pipeline.
The 4 agents are a managing_partner and 3 deal agents (deal_strategist, forensic_accountant, risk_auditor).
The deal_strategist pitches buyout theses, interprets financial data, and responds to audits.
The deal_strategist writes the high-level deal thesis `HIGH.md`.
The bod... |
pg_072__seed_1 | pg_072 | 1 | [
"lead_counsel",
"defense_strategist",
"discovery_investigator",
"shadow_prosecutor"
] | [
{
"id": "f1",
"text": "Review `THEORY.md` adversarially with the standards of a ruthless district attorney.\nYou are not part of the defense team and not their sympathetic sounding board.\nYour job is to reject narrative drift, weak alibis, and claims larger than what the admissible evidence shows.\n\nCover... | f11 | {
"lead_counsel": [
"f2"
],
"defense_strategist": [
"f7",
"f6",
"f8",
"f10",
"f9"
],
"discovery_investigator": [
"f7",
"f5",
"f3",
"f9"
],
"shadow_prosecutor": [
"f7",
"f1",
"f4",
"f9"
]
} | I need you to set up a 4-agent high-stakes criminal defense preparation pipeline.
The 4 agents are a lead_counsel and 3 trial prep agents (defense_strategist, discovery_investigator, shadow_prosecutor).
The defense_strategist plans the defense narrative, interprets evidence, and responds to mock prosecution attacks.
T... | I need you to set up a 4-agent high-stakes criminal defense preparation pipeline.
The 4 agents are a lead_counsel and 3 trial prep agents (defense_strategist, discovery_investigator, shadow_prosecutor).
The defense_strategist plans the defense narrative, interprets evidence, and responds to mock prosecution attacks.
T... |
pg_072__seed_42 | pg_072 | 42 | [
"lead_counsel",
"defense_strategist",
"discovery_investigator",
"shadow_prosecutor"
] | [
{
"id": "f1",
"text": "Score the defense viability from 1 to 10.\nRecord the verdict by overwriting the single `<mock_verdict score=X>` block at the end of `THEORY.md`.\nAlways keep exactly one `<mock_verdict>` block.\nReplace it, never append a second.\n\nScore strictly on trial merit.\nDo not give credit ... | f3 | {
"lead_counsel": [
"f4"
],
"defense_strategist": [
"f9",
"f2",
"f8",
"f7",
"f10"
],
"discovery_investigator": [
"f9",
"f5",
"f11",
"f10"
],
"shadow_prosecutor": [
"f9",
"f6",
"f1",
"f10"
]
} | I need you to set up a 4-agent high-stakes criminal defense preparation pipeline.
The 4 agents are a lead_counsel and 3 trial prep agents (defense_strategist, discovery_investigator, shadow_prosecutor).
The defense_strategist plans the defense narrative, interprets evidence, and responds to mock prosecution attacks.
T... | I need you to set up a 4-agent high-stakes criminal defense preparation pipeline.
The 4 agents are a lead_counsel and 3 trial prep agents (defense_strategist, discovery_investigator, shadow_prosecutor).
The defense_strategist plans the defense narrative, interprets evidence, and responds to mock prosecution attacks.
T... |
pg_073__seed_1 | pg_073 | 1 | [
"campaign_manager",
"chief_strategist",
"data_operative",
"opposition_researcher"
] | [
{
"id": "f1",
"text": "Review `HIGH.md` adversarially with the standards of a ruthless rival campaign.\nYou are not part of the campaign team and not their internal QA.\nYour job is to destroy the platform, expose weak evidence, and reject messaging claims larger than what the voter data actually shows.\n\n... | f11 | {
"campaign_manager": [
"f2"
],
"chief_strategist": [
"f7",
"f6",
"f8",
"f10",
"f9"
],
"data_operative": [
"f7",
"f5",
"f3",
"f9"
],
"opposition_researcher": [
"f7",
"f1",
"f4",
"f9"
]
} | I need you to set up a 4-agent campaign war room.
The 4 agents are a campaign_manager and 3 campaign agents (chief_strategist, data_operative, opposition_researcher).
The chief_strategist plans candidate messaging, interprets voter polling data, and responds to opposition attacks.
The chief_strategist writes the high-... | I need you to set up a 4-agent campaign war room.
The 4 agents are a campaign_manager and 3 campaign agents (chief_strategist, data_operative, opposition_researcher).
The chief_strategist plans candidate messaging, interprets voter polling data, and responds to opposition attacks.
The chief_strategist writes the high-... |
pg_073__seed_42 | pg_073 | 42 | [
"campaign_manager",
"chief_strategist",
"data_operative",
"opposition_researcher"
] | [
{
"id": "f1",
"text": "Score the strategy from 1 to 10.\nRecord the verdict by overwriting the single `<review score=X>` block at the end of `HIGH.md`.\nAlways keep exactly one `<review>` block.\nReplace it, never append a second.\n\nScore strictly on real electoral merit and defensibility.\nDo not give cre... | f3 | {
"campaign_manager": [
"f4"
],
"chief_strategist": [
"f9",
"f2",
"f8",
"f7",
"f10"
],
"data_operative": [
"f9",
"f5",
"f11",
"f10"
],
"opposition_researcher": [
"f9",
"f6",
"f1",
"f10"
]
} | I need you to set up a 4-agent campaign war room.
The 4 agents are a campaign_manager and 3 campaign agents (chief_strategist, data_operative, opposition_researcher).
The chief_strategist plans candidate messaging, interprets voter polling data, and responds to opposition attacks.
The chief_strategist writes the high-... | I need you to set up a 4-agent campaign war room.
The 4 agents are a campaign_manager and 3 campaign agents (chief_strategist, data_operative, opposition_researcher).
The chief_strategist plans candidate messaging, interprets voter polling data, and responds to opposition attacks.
The chief_strategist writes the high-... |
pg_074__seed_1 | pg_074 | 1 | [
"zoning_commissioner",
"chief_architect",
"civil_engineer",
"environmental_regulator"
] | [
{
"id": "f1",
"text": "Review `VISION.md` adversarially with the standards of a strict municipal authority.\nYou are not part of the development team and not their internal QA.\nYour job is to reject scope drift, weak structural evidence, and claims larger than what the traffic and load data shows, actively... | f11 | {
"zoning_commissioner": [
"f2"
],
"chief_architect": [
"f7",
"f6",
"f8",
"f10",
"f9"
],
"civil_engineer": [
"f7",
"f5",
"f3",
"f9"
],
"environmental_regulator": [
"f7",
"f1",
"f4",
"f9"
]
} | I need you to set up a 4-agent mega-infrastructure project pipeline.
The 4 agents are a zoning_commissioner and 3 project agents (chief_architect, civil_engineer, environmental_regulator).
The chief_architect plans the controversial city development vision, interprets engineering feasibility results, and responds to r... | I need you to set up a 4-agent mega-infrastructure project pipeline.
The 4 agents are a zoning_commissioner and 3 project agents (chief_architect, civil_engineer, environmental_regulator).
The chief_architect plans the controversial city development vision, interprets engineering feasibility results, and responds to r... |
pg_074__seed_42 | pg_074 | 42 | [
"zoning_commissioner",
"chief_architect",
"civil_engineer",
"environmental_regulator"
] | [
{
"id": "f1",
"text": "Score the development plan from 1 to 10.\nRecord the verdict by overwriting the single `<review score=X>` block at the end of `VISION.md`.\nAlways keep exactly one `<review>` block.\nReplace it, never append a second.\n\nScore strictly on real code compliance and environmental safety.... | f3 | {
"zoning_commissioner": [
"f4"
],
"chief_architect": [
"f9",
"f2",
"f8",
"f7",
"f10"
],
"civil_engineer": [
"f9",
"f5",
"f11",
"f10"
],
"environmental_regulator": [
"f9",
"f6",
"f1",
"f10"
]
} | I need you to set up a 4-agent mega-infrastructure project pipeline.
The 4 agents are a zoning_commissioner and 3 project agents (chief_architect, civil_engineer, environmental_regulator).
The chief_architect plans the controversial city development vision, interprets engineering feasibility results, and responds to r... | I need you to set up a 4-agent mega-infrastructure project pipeline.
The 4 agents are a zoning_commissioner and 3 project agents (chief_architect, civil_engineer, environmental_regulator).
The chief_architect plans the controversial city development vision, interprets engineering feasibility results, and responds to r... |
pg_075__seed_1 | pg_075 | 1 | [
"permit coordinator",
"structural engineer",
"field testing technician",
"city plan examiner"
] | [
{
"id": "f1",
"text": "Examine `DESIGN.md` adversarially with the standards of municipal bridge code, public safety, and permit approvability.\nYou are not part of the applicant's team and not their internal quality control.\nYour job is to reject drift, weak evidence, and claims larger than what the submit... | f11 | {
"permit coordinator": [
"f2"
],
"structural engineer": [
"f7",
"f6",
"f8",
"f10",
"f9"
],
"field testing technician": [
"f7",
"f5",
"f3",
"f9"
],
"city plan examiner": [
"f7",
"f1",
"f4",
"f9"
]
} | I need you to set up a 4-agent municipal bridge rehabilitation permitting pipeline.
The 4 agents are a permit coordinator and 3 permitting agents: structural engineer, field testing technician, city plan examiner.
The structural engineer develops the repair proposal, interprets field verification results, and responds... | I need you to set up a 4-agent municipal bridge rehabilitation permitting pipeline.
The 4 agents are a permit coordinator and 3 permitting agents: structural engineer, field testing technician, city plan examiner.
The structural engineer develops the repair proposal, interprets field verification results, and responds... |
pg_075__seed_42 | pg_075 | 42 | [
"permit coordinator",
"structural engineer",
"field testing technician",
"city plan examiner"
] | [
{
"id": "f1",
"text": "Score the work from 1 to 10.\nRecord the verdict by overwriting the single `<exam score=X>` block at the end of `DESIGN.md`.\nAlways keep exactly one `<exam>` block.\nReplace it, never append a second.\n\nScore strictly on real public-safety and permit merit.\nDo not give credit for t... | f3 | {
"permit coordinator": [
"f4"
],
"structural engineer": [
"f9",
"f2",
"f8",
"f7",
"f10"
],
"field testing technician": [
"f9",
"f5",
"f11",
"f10"
],
"city plan examiner": [
"f9",
"f6",
"f1",
"f10"
]
} | I need you to set up a 4-agent municipal bridge rehabilitation permitting pipeline.
The 4 agents are a permit coordinator and 3 permitting agents: structural engineer, field testing technician, city plan examiner.
The structural engineer develops the repair proposal, interprets field verification results, and responds... | I need you to set up a 4-agent municipal bridge rehabilitation permitting pipeline.
The 4 agents are a permit coordinator and 3 permitting agents: structural engineer, field testing technician, city plan examiner.
The structural engineer develops the repair proposal, interprets field verification results, and responds... |
pg_076__seed_1 | pg_076 | 1 | [
"case manager",
"appellate attorney",
"legal research paralegal",
"moot-court judge"
] | [
{
"id": "f1",
"text": "Review `HIGH.md` adversarially with the standards of a demanding appellate panel.\nYou are not part of the litigation team and not their internal cite-checker.\nYour job is to reject drift, weak support, and arguments larger than what the authorities and record show.\n\nCover four lay... | f11 | {
"case manager": [
"f2"
],
"appellate attorney": [
"f7",
"f6",
"f8",
"f10",
"f9"
],
"legal research paralegal": [
"f7",
"f5",
"f3",
"f9"
],
"moot-court judge": [
"f7",
"f1",
"f4",
"f9"
]
} | I need you to set up a 4-agent appellate litigation preparation pipeline.
The 4 agents are a case manager and 3 legal preparation agents (appellate attorney, legal research paralegal, moot-court judge).
The appellate attorney builds the appellate theory, interprets the record and authorities, and responds to moot-cour... | I need you to set up a 4-agent appellate litigation preparation pipeline.
The 4 agents are a case manager and 3 legal preparation agents (appellate attorney, legal research paralegal, moot-court judge).
The appellate attorney builds the appellate theory, interprets the record and authorities, and responds to moot-cour... |
pg_076__seed_42 | pg_076 | 42 | [
"case manager",
"appellate attorney",
"legal research paralegal",
"moot-court judge"
] | [
{
"id": "f1",
"text": "Score the work from 1 to 10.\nRecord the verdict by overwriting the single `<review score=X>` block at the end of `HIGH.md`.\nAlways keep exactly one `<review>` block.\nReplace it, never append a second.\n\nScore strictly on real appellate merit.\nDo not give credit for the work being... | f3 | {
"case manager": [
"f4"
],
"appellate attorney": [
"f9",
"f2",
"f8",
"f7",
"f10"
],
"legal research paralegal": [
"f9",
"f5",
"f11",
"f10"
],
"moot-court judge": [
"f9",
"f6",
"f1",
"f10"
]
} | I need you to set up a 4-agent appellate litigation preparation pipeline.
The 4 agents are a case manager and 3 legal preparation agents (appellate attorney, legal research paralegal, moot-court judge).
The appellate attorney builds the appellate theory, interprets the record and authorities, and responds to moot-cour... | I need you to set up a 4-agent appellate litigation preparation pipeline.
The 4 agents are a case manager and 3 legal preparation agents (appellate attorney, legal research paralegal, moot-court judge).
The appellate attorney builds the appellate theory, interprets the record and authorities, and responds to moot-cour... |
pg_077__seed_1 | pg_077 | 1 | [
"acquisitions coordinator",
"curator",
"collections registrar",
"provenance auditor"
] | [
{
"id": "f1",
"text": "Review `ACQ.md` adversarially with the standards of museum ethics, cultural-property law, and responsible acquisition practice.\nYou are not part of the acquisition team and not their internal quality check.\nYour job is to reject drift, weak documentation, and acquisition claims larg... | f11 | {
"acquisitions coordinator": [
"f2"
],
"curator": [
"f7",
"f6",
"f8",
"f10",
"f9"
],
"collections registrar": [
"f7",
"f5",
"f3",
"f9"
],
"provenance auditor": [
"f7",
"f1",
"f4",
"f9"
]
} | I need you to set up a 4-agent museum acquisition due-diligence pipeline.
The 4 agents are an acquisitions coordinator and 3 acquisition-review agents: curator, collections registrar, and provenance auditor.
The curator builds the acquisition case, interprets provenance findings, and responds to ethics review.
The cur... | I need you to set up a 4-agent museum acquisition due-diligence pipeline.
The 4 agents are an acquisitions coordinator and 3 acquisition-review agents: curator, collections registrar, and provenance auditor.
The curator builds the acquisition case, interprets provenance findings, and responds to ethics review.
The cur... |
pg_077__seed_42 | pg_077 | 42 | [
"acquisitions coordinator",
"curator",
"collections registrar",
"provenance auditor"
] | [
{
"id": "f1",
"text": "Score the work from 1 to 10.\nRecord the verdict by overwriting the single `<review score=X>` block at the end of `ACQ.md`.\nAlways keep exactly one `<review>` block.\nReplace it, never append a second.\n\nScore strictly on real merit.\nDo not give credit for the acquisition file bein... | f3 | {
"acquisitions coordinator": [
"f4"
],
"curator": [
"f9",
"f2",
"f8",
"f7",
"f10"
],
"collections registrar": [
"f9",
"f5",
"f11",
"f10"
],
"provenance auditor": [
"f9",
"f6",
"f1",
"f10"
]
} | I need you to set up a 4-agent museum acquisition due-diligence pipeline.
The 4 agents are an acquisitions coordinator and 3 acquisition-review agents: curator, collections registrar, and provenance auditor.
The curator builds the acquisition case, interprets provenance findings, and responds to ethics review.
The cur... | I need you to set up a 4-agent museum acquisition due-diligence pipeline.
The 4 agents are an acquisitions coordinator and 3 acquisition-review agents: curator, collections registrar, and provenance auditor.
The curator builds the acquisition case, interprets provenance findings, and responds to ethics review.
The cur... |
pg_078__seed_1 | pg_078 | 1 | [
"service coordinator",
"executive chef",
"sous chef",
"restaurant critic"
] | [
{
"id": "f1",
"text": "Critique `MENU.md` adversarially with the standards of a top restaurant assessment.\nYou are not part of the team and not their internal quality-control cook.\nYour job is to reject drift, weak evidence, and claims larger than what the actual plates show.\n\nCover four layers.\n\n- Dr... | f11 | {
"service coordinator": [
"f2"
],
"executive chef": [
"f7",
"f6",
"f8",
"f10",
"f9"
],
"sous chef": [
"f7",
"f5",
"f3",
"f9"
],
"restaurant critic": [
"f7",
"f1",
"f4",
"f9"
]
} | I need you to set up a 4-agent fine-dining menu launch validation pipeline.
The 4 agents are a service coordinator and 3 launch agents: executive chef, sous chef, and restaurant critic.
The executive chef develops the tasting-menu concept, plans service trials, interprets guest and kitchen results, and responds to cri... | I need you to set up a 4-agent fine-dining menu launch validation pipeline.
The 4 agents are a service coordinator and 3 launch agents: executive chef, sous chef, and restaurant critic.
The executive chef develops the tasting-menu concept, plans service trials, interprets guest and kitchen results, and responds to cri... |
pg_078__seed_42 | pg_078 | 42 | [
"service coordinator",
"executive chef",
"sous chef",
"restaurant critic"
] | [
{
"id": "f1",
"text": "Score the work from 1 to 10.\nRecord the verdict by overwriting the single `<critic score=X>` block at the end of `MENU.md`.\nAlways keep exactly one `<critic>` block.\nReplace it, never append a second.\n\nScore strictly on real merit.\nDo not give credit for the work being honest ab... | f3 | {
"service coordinator": [
"f4"
],
"executive chef": [
"f9",
"f2",
"f8",
"f7",
"f10"
],
"sous chef": [
"f9",
"f5",
"f11",
"f10"
],
"restaurant critic": [
"f9",
"f6",
"f1",
"f10"
]
} | I need you to set up a 4-agent fine-dining menu launch validation pipeline.
The 4 agents are a service coordinator and 3 launch agents: executive chef, sous chef, and restaurant critic.
The executive chef develops the tasting-menu concept, plans service trials, interprets guest and kitchen results, and responds to cri... | I need you to set up a 4-agent fine-dining menu launch validation pipeline.
The 4 agents are a service coordinator and 3 launch agents: executive chef, sous chef, and restaurant critic.
The executive chef develops the tasting-menu concept, plans service trials, interprets guest and kitchen results, and responds to cri... |
pg_079__seed_1 | pg_079 | 1 | [
"drill coordinator",
"life safety planner",
"evacuation drill lead",
"fire marshal"
] | [
{
"id": "f1",
"text": "Review `EVAC_PLAN.md` adversarially with the standards of an authority having jurisdiction certifying a large public venue.\nYou are not part of the venue team and not their internal preparedness coach.\nYour job is to reject drift, weak evidence, and claims larger than what the drill... | f11 | {
"drill coordinator": [
"f2"
],
"life safety planner": [
"f7",
"f6",
"f8",
"f10",
"f9"
],
"evacuation drill lead": [
"f7",
"f5",
"f3",
"f9"
],
"fire marshal": [
"f7",
"f1",
"f4",
"f9"
]
} | I need you to set up a 4-agent evacuation certification pipeline.
The 4 agents are a drill coordinator and 3 certification agents: life safety planner, evacuation drill lead, fire marshal.
The life safety planner proposes the evacuation strategy, interprets drill results, and responds to compliance challenges.
The lif... | I need you to set up a 4-agent evacuation certification pipeline.
The 4 agents are a drill coordinator and 3 certification agents: life safety planner, evacuation drill lead, fire marshal.
The life safety planner proposes the evacuation strategy, interprets drill results, and responds to compliance challenges.
The lif... |
pg_079__seed_42 | pg_079 | 42 | [
"drill coordinator",
"life safety planner",
"evacuation drill lead",
"fire marshal"
] | [
{
"id": "f1",
"text": "Score the work from 1 to 10.\nRecord the verdict by overwriting the single `<review score=X>` block at the end of `EVAC_PLAN.md`.\nAlways keep exactly one `<review>` block.\nReplace it, never append a second.\n\nScore strictly on real certification merit.\nDo not give credit for the w... | f3 | {
"drill coordinator": [
"f4"
],
"life safety planner": [
"f9",
"f2",
"f8",
"f7",
"f10"
],
"evacuation drill lead": [
"f9",
"f5",
"f11",
"f10"
],
"fire marshal": [
"f9",
"f6",
"f1",
"f10"
]
} | I need you to set up a 4-agent evacuation certification pipeline.
The 4 agents are a drill coordinator and 3 certification agents: life safety planner, evacuation drill lead, fire marshal.
The life safety planner proposes the evacuation strategy, interprets drill results, and responds to compliance challenges.
The lif... | I need you to set up a 4-agent evacuation certification pipeline.
The 4 agents are a drill coordinator and 3 certification agents: life safety planner, evacuation drill lead, fire marshal.
The life safety planner proposes the evacuation strategy, interprets drill results, and responds to compliance challenges.
The lif... |
pg_080__seed_1 | pg_080 | 1 | [
"secops-director",
"threat-modeler",
"security-architect",
"red-teamer",
"vulnerability-analyst"
] | [
{
"id": "f1",
"text": "Keep `SOLUTION.md` lean. Its body — everything except the `<review>` block — must stay under 3000 words (~20KB), about a four-page note; record only the final exploitation state and move iteration-by-iteration history into a separate `CHANGELOG.md`. Bold text must not exceed 5% of the... | f10 | {
"secops-director": [
"f2"
],
"threat-modeler": [
"f3",
"f5",
"f7",
"f11",
"f6"
],
"security-architect": [
"f3",
"f7",
"f12",
"f4",
"f6"
],
"red-teamer": [
"f3",
"f8",
"f1",
"f13",
"f6"
],
"vulnerability-analyst": [
"f3",
"f9",
... | I need you to set up a 5-agent cybersecurity pipeline that assesses target systems by developing abstract threat models and concrete red-team exploitation scripts in parallel, with each side informing the other. A system assessment request arrives as a natural-language description in `problem.md`. The pipeline alternat... | I need you to set up a 5-agent cybersecurity pipeline that assesses target systems by developing abstract threat models and concrete red-team exploitation scripts in parallel, with each side informing the other. A system assessment request arrives as a natural-language description in `problem.md`. The pipeline alternat... |
pg_080__seed_42 | pg_080 | 42 | [
"secops-director",
"threat-modeler",
"security-architect",
"red-teamer",
"vulnerability-analyst"
] | [
{
"id": "f1",
"text": "Read `problem.md` and `THEORY.md`, then implement the red-team exploit scripts:\n\n- Include every import. Create each payload with the correct encoding, define the network targets and memory offsets exactly as the threat model outlines. Execute the script, check the response status, ... | f10 | {
"secops-director": [
"f5"
],
"threat-modeler": [
"f11",
"f6",
"f2",
"f12",
"f9"
],
"security-architect": [
"f11",
"f2",
"f4",
"f7",
"f9"
],
"red-teamer": [
"f11",
"f1",
"f8",
"f3",
"f9"
],
"vulnerability-analyst": [
"f11",
"f13"... | I need you to set up a 5-agent cybersecurity pipeline that assesses target systems by developing abstract threat models and concrete red-team exploitation scripts in parallel, with each side informing the other. A system assessment request arrives as a natural-language description in `problem.md`. The pipeline alternat... | I need you to set up a 5-agent cybersecurity pipeline that assesses target systems by developing abstract threat models and concrete red-team exploitation scripts in parallel, with each side informing the other. A system assessment request arrives as a natural-language description in `problem.md`. The pipeline alternat... |
pg_081__seed_1 | pg_081 | 1 | [
"chief-engineer",
"aerodynamicist",
"flight-dynamics-evaluator",
"cfd-modeler",
"telemetry-auditor"
] | [
{
"id": "f1",
"text": "Keep `SOLUTION.md` lean. Its body — everything except the `<review>` block — must stay under 3000 words (~20KB), about a four-page note; record only the final mesh configuration, solver setup, and converged results, moving iteration-by-iteration residual history into a separate `CHANG... | f10 | {
"chief-engineer": [
"f2"
],
"aerodynamicist": [
"f3",
"f5",
"f7",
"f11",
"f6"
],
"flight-dynamics-evaluator": [
"f3",
"f7",
"f12",
"f4",
"f6"
],
"cfd-modeler": [
"f3",
"f8",
"f1",
"f13",
"f6"
],
"telemetry-auditor": [
"f3",
"f9"... | I need you to set up a 5-agent pipeline that designs aerospace vehicles by alternating between mathematical aerodynamic calculations of performance bounds and computational fluid dynamics (CFD) physical simulations, with each side informing the other. A mission profile and design requirements arrive as a natural-langua... | I need you to set up a 5-agent pipeline that designs aerospace vehicles by alternating between mathematical aerodynamic calculations of performance bounds and computational fluid dynamics (CFD) physical simulations, with each side informing the other. A mission profile and design requirements arrive as a natural-langua... |
pg_081__seed_42 | pg_081 | 42 | [
"chief-engineer",
"aerodynamicist",
"flight-dynamics-evaluator",
"cfd-modeler",
"telemetry-auditor"
] | [
{
"id": "f1",
"text": "Read `problem.md` and `THEORY.md`, then implement the CFD simulation script:\n\n- Include every import (e.g., meshing libraries, CFD solver python APIs like SU2 or OpenFOAM wrappers). Generate the mesh with the correct dimensions and boundary conditions, define the fluid properties, a... | f10 | {
"chief-engineer": [
"f5"
],
"aerodynamicist": [
"f11",
"f6",
"f2",
"f12",
"f9"
],
"flight-dynamics-evaluator": [
"f11",
"f2",
"f4",
"f7",
"f9"
],
"cfd-modeler": [
"f11",
"f1",
"f8",
"f3",
"f9"
],
"telemetry-auditor": [
"f11",
"f... | I need you to set up a 5-agent pipeline that designs aerospace vehicles by alternating between mathematical aerodynamic calculations of performance bounds and computational fluid dynamics (CFD) physical simulations, with each side informing the other. A mission profile and design requirements arrive as a natural-langua... | I need you to set up a 5-agent pipeline that designs aerospace vehicles by alternating between mathematical aerodynamic calculations of performance bounds and computational fluid dynamics (CFD) physical simulations, with each side informing the other. A mission profile and design requirements arrive as a natural-langua... |
pg_082__seed_1 | pg_082 | 1 | [
"lead-producer",
"economy-designer",
"meta-strategist",
"bot-scripter",
"exploit-investigator"
] | [
{
"id": "f1",
"text": "Keep `PLAYTEST.md` lean. Its body — everything except the `<review>` block — must stay under 3000 words (~20KB), about a four-page note; record only the final state and move iteration-by-iteration history into a separate `CHANGELOG.md`. Bold text must not exceed 5% of the file: if eve... | f10 | {
"lead-producer": [
"f2"
],
"economy-designer": [
"f3",
"f5",
"f7",
"f11",
"f6"
],
"meta-strategist": [
"f3",
"f7",
"f12",
"f4",
"f6"
],
"bot-scripter": [
"f3",
"f8",
"f1",
"f13",
"f6"
],
"exploit-investigator": [
"f3",
"f9",
... | I need you to set up a 5-agent pipeline that solves multiplayer video game economy balancing by developing spreadsheet-based macroeconomic limits and automated bot playtesting in parallel, with each side informing the other. A problem arrives as a natural-language statement in `problem.md`. The pipeline alternates betw... | I need you to set up a 5-agent pipeline that solves multiplayer video game economy balancing by developing spreadsheet-based macroeconomic limits and automated bot playtesting in parallel, with each side informing the other. A problem arrives as a natural-language statement in `problem.md`. The pipeline alternates betw... |
pg_082__seed_42 | pg_082 | 42 | [
"lead-producer",
"economy-designer",
"meta-strategist",
"bot-scripter",
"exploit-investigator"
] | [
{
"id": "f1",
"text": "Read `problem.md` and `MACRO.md`, then implement the automated playtest bot:\n\n- Include every import. Create each state variable with the correct type and bounds, define the bot's action space and every game rule exactly as the problem states. Run the simulation, check the bot's fin... | f10 | {
"lead-producer": [
"f5"
],
"economy-designer": [
"f11",
"f6",
"f2",
"f12",
"f9"
],
"meta-strategist": [
"f11",
"f2",
"f4",
"f7",
"f9"
],
"bot-scripter": [
"f11",
"f1",
"f8",
"f3",
"f9"
],
"exploit-investigator": [
"f11",
"f13",
... | I need you to set up a 5-agent pipeline that solves multiplayer video game economy balancing by developing spreadsheet-based macroeconomic limits and automated bot playtesting in parallel, with each side informing the other. A problem arrives as a natural-language statement in `problem.md`. The pipeline alternates betw... | I need you to set up a 5-agent pipeline that solves multiplayer video game economy balancing by developing spreadsheet-based macroeconomic limits and automated bot playtesting in parallel, with each side informing the other. A problem arrives as a natural-language statement in `problem.md`. The pipeline alternates betw... |
pg_083__seed_1 | pg_083 | 1 | [
"research-director",
"computational-biologist",
"pharmacokinetics-assessor",
"synthesis-chemist",
"assay-validator"
] | [
{
"id": "f1",
"text": "Keep `IN_VITRO.md` lean. Its body — everything except the `<review>` block — must stay under 3000 words (~20KB), about a four-page summary; record only the final optimized route and assay readouts, and move iteration-by-iteration failed reaction history into a separate `CHANGELOG.md`.... | f10 | {
"research-director": [
"f2"
],
"computational-biologist": [
"f3",
"f5",
"f7",
"f11",
"f6"
],
"pharmacokinetics-assessor": [
"f3",
"f7",
"f12",
"f4",
"f6"
],
"synthesis-chemist": [
"f3",
"f8",
"f1",
"f13",
"f6"
],
"assay-validator": [
... | I need you to set up a 5-agent pipeline for preclinical drug discovery that develops in-silico molecular docking models and empirical in-vitro synthesis and assay testing in parallel, with each side informing the other. A target disease profile and desired therapeutic mechanism arrive as a natural-language statement in... | I need you to set up a 5-agent pipeline for preclinical drug discovery that develops in-silico molecular docking models and empirical in-vitro synthesis and assay testing in parallel, with each side informing the other. A target disease profile and desired therapeutic mechanism arrive as a natural-language statement in... |
pg_083__seed_42 | pg_083 | 42 | [
"research-director",
"computational-biologist",
"pharmacokinetics-assessor",
"synthesis-chemist",
"assay-validator"
] | [
{
"id": "f1",
"text": "Read `target_profile.md` and `IN_SILICO.md`, then execute the synthesis and bioassays:\n\n- Include every required reagent and solvent. Detail the retrosynthetic route, reaction conditions, and purification steps. Execute the synthesis, run the biochemical and evaluationular assays ex... | f10 | {
"research-director": [
"f5"
],
"computational-biologist": [
"f11",
"f6",
"f2",
"f12",
"f9"
],
"pharmacokinetics-assessor": [
"f11",
"f2",
"f4",
"f7",
"f9"
],
"synthesis-chemist": [
"f11",
"f1",
"f8",
"f3",
"f9"
],
"assay-validator": [
... | I need you to set up a 5-agent pipeline for preclinical drug discovery that develops in-silico molecular docking models and empirical in-vitro synthesis and assay testing in parallel, with each side informing the other. A target disease profile and desired therapeutic mechanism arrive as a natural-language statement in... | I need you to set up a 5-agent pipeline for preclinical drug discovery that develops in-silico molecular docking models and empirical in-vitro synthesis and assay testing in parallel, with each side informing the other. A target disease profile and desired therapeutic mechanism arrive as a natural-language statement in... |
pg_084__seed_1 | pg_084 | 1 | [
"drafting-chair",
"macroeconomist",
"tax-attorney",
"tax-strategist",
"irs-auditor"
] | [
{
"id": "f1",
"text": "Keep `EVASION.md` lean. Its body — everything except the `<review>` block — must stay under 3000 words (~20KB), about a four-page brief; record only the final corporate structure and move iteration-by-iteration history into a separate `CHANGELOG.md`. Bold text must not exceed 5% of th... | f10 | {
"drafting-chair": [
"f2"
],
"macroeconomist": [
"f3",
"f5",
"f7",
"f11",
"f6"
],
"tax-attorney": [
"f3",
"f7",
"f12",
"f4",
"f6"
],
"tax-strategist": [
"f3",
"f8",
"f1",
"f13",
"f6"
],
"irs-auditor": [
"f3",
"f9",
"f4",
... | I need you to set up a 5-agent pipeline that develops corporate tax legislation by drafting macroeconomic policy and running adversarial accounting simulations in parallel, with each side informing the other. A legislative mandate arrives as a natural-language statement in `problem.md`. The pipeline alternates between ... | I need you to set up a 5-agent pipeline that develops corporate tax legislation by drafting macroeconomic policy and running adversarial accounting simulations in parallel, with each side informing the other. A legislative mandate arrives as a natural-language statement in `problem.md`. The pipeline alternates between ... |
pg_084__seed_42 | pg_084 | 42 | [
"drafting-chair",
"macroeconomist",
"tax-attorney",
"tax-strategist",
"irs-auditor"
] | [
{
"id": "f1",
"text": "Read `problem.md` and `POLICY.md`, then implement the adversarial accounting simulation:\n\n- Include every import. Create each financial variable and subsidiary with the correct jurisdiction and capitalization, define the tax minimization objective and every legal constraint exactly ... | f10 | {
"drafting-chair": [
"f5"
],
"macroeconomist": [
"f11",
"f6",
"f2",
"f12",
"f9"
],
"tax-attorney": [
"f11",
"f2",
"f4",
"f7",
"f9"
],
"tax-strategist": [
"f11",
"f1",
"f8",
"f3",
"f9"
],
"irs-auditor": [
"f11",
"f13",
"f7",
... | I need you to set up a 5-agent pipeline that develops corporate tax legislation by drafting macroeconomic policy and running adversarial accounting simulations in parallel, with each side informing the other. A legislative mandate arrives as a natural-language statement in `problem.md`. The pipeline alternates between ... | I need you to set up a 5-agent pipeline that develops corporate tax legislation by drafting macroeconomic policy and running adversarial accounting simulations in parallel, with each side informing the other. A legislative mandate arrives as a natural-language statement in `problem.md`. The pipeline alternates between ... |
pg_085__seed_1 | pg_085 | 1 | [
"trial operations coordinator",
"clinical protocol architect",
"independent scientific and ethics reviewer",
"statistical analysis plan author",
"regulatory methods auditor"
] | [
{
"id": "f1",
"text": "Keep `SAP.md` lean. Its body — everything except the `<review>` block — must stay under 3000 words (~20KB), about a four-page note; record only the final state and move iteration-by-iteration history into a separate `CHANGELOG.md`. Bold text must not exceed 5% of the file: if everythi... | f10 | {
"trial operations coordinator": [
"f2"
],
"clinical protocol architect": [
"f3",
"f5",
"f7",
"f11",
"f6"
],
"independent scientific and ethics reviewer": [
"f3",
"f7",
"f12",
"f4",
"f6"
],
"statistical analysis plan author": [
"f3",
"f8",
"f1",
... | I need you to set up a 5-agent pipeline that develops a clinical trial protocol by evolving the scientific trial rationale and the statistical analysis plan in parallel, with each side informing the other under skeptical ethics, methods, and regulatory review. A trial concept arrives as a natural-language statement in ... | I need you to set up a 5-agent pipeline that develops a clinical trial protocol by evolving the scientific trial rationale and the statistical analysis plan in parallel, with each side informing the other under skeptical ethics, methods, and regulatory review. A trial concept arrives as a natural-language statement in ... |
pg_085__seed_42 | pg_085 | 42 | [
"trial operations coordinator",
"clinical protocol architect",
"independent scientific and ethics reviewer",
"statistical analysis plan author",
"regulatory methods auditor"
] | [
{
"id": "f1",
"text": "Read `trial_brief.md` and `PROTOCOL_RATIONALE.md`, then produce the statistical analysis plan:\n\n- Include every essential design and analysis element. Define the estimands, endpoints, analysis populations, hypotheses, type I error strategy, sample-size assumptions, randomization sch... | f10 | {
"trial operations coordinator": [
"f5"
],
"clinical protocol architect": [
"f11",
"f6",
"f2",
"f12",
"f9"
],
"independent scientific and ethics reviewer": [
"f11",
"f2",
"f4",
"f7",
"f9"
],
"statistical analysis plan author": [
"f11",
"f1",
"f8",
... | I need you to set up a 5-agent pipeline that develops a clinical trial protocol by evolving the scientific trial rationale and the statistical analysis plan in parallel, with each side informing the other under skeptical ethics, methods, and regulatory review. A trial concept arrives as a natural-language statement in ... | I need you to set up a 5-agent pipeline that develops a clinical trial protocol by evolving the scientific trial rationale and the statistical analysis plan in parallel, with each side informing the other under skeptical ethics, methods, and regulatory review. A trial concept arrives as a natural-language statement in ... |
pg_086__seed_1 | pg_086 | 1 | [
"litigation workflow coordinator",
"case theory counsel",
"devil's-advocate legal reviewer",
"appellate brief drafter",
"opposing-counsel brief examiner"
] | [
{
"id": "f1",
"text": "Keep `BRIEF.md` lean. Its body — everything except the `<review>` block — must stay under 3000 words (~20KB), about a compact internal appellate draft; record only the current filing-ready position and move iteration-by-iteration history into a separate `CHANGELOG.md`. Bold text must ... | f10 | {
"litigation workflow coordinator": [
"f2"
],
"case theory counsel": [
"f3",
"f5",
"f7",
"f11",
"f6"
],
"devil's-advocate legal reviewer": [
"f3",
"f7",
"f12",
"f4",
"f6"
],
"appellate brief drafter": [
"f3",
"f8",
"f1",
"f13",
"f6"
],
"... | I need you to set up a 5-agent pipeline for appellate litigation preparation. The team develops a case theory and a draft appellate brief in parallel, with each side informing the other. The case materials arrive as a record-and-issues packet in `record.md`. The pipeline alternates between revising the case theory (sta... | I need you to set up a 5-agent pipeline for appellate litigation preparation. The team develops a case theory and a draft appellate brief in parallel, with each side informing the other. The case materials arrive as a record-and-issues packet in `record.md`. The pipeline alternates between revising the case theory (sta... |
pg_086__seed_42 | pg_086 | 42 | [
"litigation workflow coordinator",
"case theory counsel",
"devil's-advocate legal reviewer",
"appellate brief drafter",
"opposing-counsel brief examiner"
] | [
{
"id": "f1",
"text": "Read `record.md` and `CASE_THEORY.md`, then draft the appellate brief:\n\n- Include the required appellate components appropriate to the matter: jurisdictional statement, issues presented, statement of the case, statement of facts, summary of argument, argument headings, standard of r... | f10 | {
"litigation workflow coordinator": [
"f5"
],
"case theory counsel": [
"f11",
"f6",
"f2",
"f12",
"f9"
],
"devil's-advocate legal reviewer": [
"f11",
"f2",
"f4",
"f7",
"f9"
],
"appellate brief drafter": [
"f11",
"f1",
"f8",
"f3",
"f9"
],
... | I need you to set up a 5-agent pipeline for appellate litigation preparation. The team develops a case theory and a draft appellate brief in parallel, with each side informing the other. The case materials arrive as a record-and-issues packet in `record.md`. The pipeline alternates between revising the case theory (sta... | I need you to set up a 5-agent pipeline for appellate litigation preparation. The team develops a case theory and a draft appellate brief in parallel, with each side informing the other. The case materials arrive as a record-and-issues packet in `record.md`. The pipeline alternates between revising the case theory (sta... |
pg_087__seed_1 | pg_087 | 1 | [
"authentication case manager",
"provenance researcher",
"provenance skeptic",
"conservation science examiner",
"forensic methods auditor"
] | [
{
"id": "f1",
"text": "Keep `FORENSICS.md` lean. Its body — everything except the `<review>` block — must stay under 3000 words (~20KB), about a four-page note; record only the final state and move iteration-by-iteration history, raw instrument logs, and extended spectra tables into a separate `CHANGELOG.md... | f10 | {
"authentication case manager": [
"f2"
],
"provenance researcher": [
"f3",
"f5",
"f7",
"f11",
"f6"
],
"provenance skeptic": [
"f3",
"f7",
"f12",
"f4",
"f6"
],
"conservation science examiner": [
"f3",
"f8",
"f1",
"f13",
"f6"
],
"forensic ... | I need you to set up a 5-agent pipeline that evaluates a fine-art authentication case by developing a provenance hypothesis and a forensic examination report in parallel, with each side informing the other. A case arrives as a natural-language brief in `case.md`. The pipeline alternates between revising the provenance ... | I need you to set up a 5-agent pipeline that evaluates a fine-art authentication case by developing a provenance hypothesis and a forensic examination report in parallel, with each side informing the other. A case arrives as a natural-language brief in `case.md`. The pipeline alternates between revising the provenance ... |
pg_087__seed_42 | pg_087 | 42 | [
"authentication case manager",
"provenance researcher",
"provenance skeptic",
"conservation science examiner",
"forensic methods auditor"
] | [
{
"id": "f1",
"text": "Read `case.md` and `PROVENANCE.md`, then produce the conservation-science examination:\n\n- Include every relevant method used or recommended. Document imaging, microscopy, UV/IR/visible observations, X-radiography, XRF, Raman, FTIR, GC-MS, dendrochronology, radiocarbon, fiber analysi... | f10 | {
"authentication case manager": [
"f5"
],
"provenance researcher": [
"f11",
"f6",
"f2",
"f12",
"f9"
],
"provenance skeptic": [
"f11",
"f2",
"f4",
"f7",
"f9"
],
"conservation science examiner": [
"f11",
"f1",
"f8",
"f3",
"f9"
],
"forensic... | I need you to set up a 5-agent pipeline that evaluates a fine-art authentication case by developing a provenance hypothesis and a forensic examination report in parallel, with each side informing the other. A case arrives as a natural-language brief in `case.md`. The pipeline alternates between revising the provenance ... | I need you to set up a 5-agent pipeline that evaluates a fine-art authentication case by developing a provenance hypothesis and a forensic examination report in parallel, with each side informing the other. A case arrives as a natural-language brief in `case.md`. The pipeline alternates between revising the provenance ... |
pg_088__seed_1 | pg_088 | 1 | [
"project coordination engineer",
"structural assessment engineer",
"independent calculation checker",
"retrofit design engineer",
"building-code plan examiner"
] | [
{
"id": "f1",
"text": "Keep `RETROFIT.md` lean. Its body — everything except the `<review>` block — must stay under 3000 words (~20KB), about a four-page engineering note; record only the final state and move iteration-by-iteration history into a separate `CHANGELOG.md`. Bold text must not exceed 5% of the ... | f10 | {
"project coordination engineer": [
"f2"
],
"structural assessment engineer": [
"f3",
"f5",
"f7",
"f11",
"f6"
],
"independent calculation checker": [
"f3",
"f7",
"f12",
"f4",
"f6"
],
"retrofit design engineer": [
"f3",
"f8",
"f1",
"f13",
"f6... | I need you to set up a 5-agent pipeline for bridge retrofit engineering. The work alternates between a structural safety assessment and a retrofit design package, with each side informing the other. A bridge condition brief arrives as a natural-language statement in `bridge_brief.md`. The pipeline alternates between re... | I need you to set up a 5-agent pipeline for bridge retrofit engineering. The work alternates between a structural safety assessment and a retrofit design package, with each side informing the other. A bridge condition brief arrives as a natural-language statement in `bridge_brief.md`. The pipeline alternates between re... |
pg_088__seed_42 | pg_088 | 42 | [
"project coordination engineer",
"structural assessment engineer",
"independent calculation checker",
"retrofit design engineer",
"building-code plan examiner"
] | [
{
"id": "f1",
"text": "Read `bridge_brief.md` and `ASSESSMENT.md`, then prepare the retrofit design package:\n\n- Include the governing design assumptions, material specifications, member sizes, connection details, anchor layouts, construction staging constraints, and any required calculations. Define the r... | f10 | {
"project coordination engineer": [
"f5"
],
"structural assessment engineer": [
"f11",
"f6",
"f2",
"f12",
"f9"
],
"independent calculation checker": [
"f11",
"f2",
"f4",
"f7",
"f9"
],
"retrofit design engineer": [
"f11",
"f1",
"f8",
"f3",
"f... | I need you to set up a 5-agent pipeline for bridge retrofit engineering. The work alternates between a structural safety assessment and a retrofit design package, with each side informing the other. A bridge condition brief arrives as a natural-language statement in `bridge_brief.md`. The pipeline alternates between re... | I need you to set up a 5-agent pipeline for bridge retrofit engineering. The work alternates between a structural safety assessment and a retrofit design package, with each side informing the other. A bridge condition brief arrives as a natural-language statement in `bridge_brief.md`. The pipeline alternates between re... |
pg_089__seed_1 | pg_089 | 1 | [
"incident coordination officer",
"fire behavior analyst",
"red-team fire model reviewer",
"operations plan chief",
"incident safety officer"
] | [
{
"id": "f1",
"text": "Keep `IAP.md` lean. Its body — everything except the `<review>` block — must stay under 3000 words (~20KB), about a four-page operational note; record iteration-by-iteration history in a separate `CHANGELOG.md`. Bold text must not exceed 5% of the file: if everything is bold, nothing ... | f10 | {
"incident coordination officer": [
"f2"
],
"fire behavior analyst": [
"f3",
"f5",
"f7",
"f11",
"f6"
],
"red-team fire model reviewer": [
"f3",
"f7",
"f12",
"f4",
"f6"
],
"operations plan chief": [
"f3",
"f8",
"f1",
"f13",
"f6"
],
"incid... | I need you to set up a 5-agent pipeline that develops a wildfire incident response plan by refining fire-behavior forecasts and tactical operations planning in parallel, with each side informing the other. An incident package arrives as a natural-language briefing in `incident.md`. The pipeline alternates between revis... | I need you to set up a 5-agent pipeline that develops a wildfire incident response plan by refining fire-behavior forecasts and tactical operations planning in parallel, with each side informing the other. An incident package arrives as a natural-language briefing in `incident.md`. The pipeline alternates between revis... |
pg_089__seed_42 | pg_089 | 42 | [
"incident coordination officer",
"fire behavior analyst",
"red-team fire model reviewer",
"operations plan chief",
"incident safety officer"
] | [
{
"id": "f1",
"text": "Read `incident.md` and `FORECAST.md`, then produce the tactical operations plan:\n\n- Include the incident objectives, operational period, division or group assignments, available resources, communications assumptions, medical plan references, evacuation coordination needs, contingenc... | f10 | {
"incident coordination officer": [
"f5"
],
"fire behavior analyst": [
"f11",
"f6",
"f2",
"f12",
"f9"
],
"red-team fire model reviewer": [
"f11",
"f2",
"f4",
"f7",
"f9"
],
"operations plan chief": [
"f11",
"f1",
"f8",
"f3",
"f9"
],
"inci... | I need you to set up a 5-agent pipeline that develops a wildfire incident response plan by refining fire-behavior forecasts and tactical operations planning in parallel, with each side informing the other. An incident package arrives as a natural-language briefing in `incident.md`. The pipeline alternates between revis... | I need you to set up a 5-agent pipeline that develops a wildfire incident response plan by refining fire-behavior forecasts and tactical operations planning in parallel, with each side informing the other. An incident package arrives as a natural-language briefing in `incident.md`. The pipeline alternates between revis... |
pg_090__seed_1 | pg_090 | 1 | [
"litigation-manager",
"legal-strategist",
"shadow-counsel",
"discovery-specialist",
"evidentiary-auditor"
] | [
{
"id": "f1",
"text": "Keep `DISCOVERY.md` lean. Its body — everything except the `<review>` block — must stay under 3000 words (~20KB), about a four-page memo; record only the current synthesized evidentiary record and move iteration-by-iteration search history into a separate `CHANGELOG.md`. Bold text mus... | f10 | {
"litigation-manager": [
"f2"
],
"legal-strategist": [
"f3",
"f5",
"f7",
"f11",
"f6"
],
"shadow-counsel": [
"f3",
"f7",
"f12",
"f4",
"f6"
],
"discovery-specialist": [
"f3",
"f8",
"f1",
"f13",
"f6"
],
"evidentiary-auditor": [
"f3",
... | I need you to set up a 5-agent pipeline that manages corporate litigation defense by developing abstract legal argumentation strategy and concrete evidentiary discovery in parallel, with each side informing the other. A lawsuit arrives as a natural-language statement in `lawsuit.md`. The pipeline alternates between rev... | I need you to set up a 5-agent pipeline that manages corporate litigation defense by developing abstract legal argumentation strategy and concrete evidentiary discovery in parallel, with each side informing the other. A lawsuit arrives as a natural-language statement in `lawsuit.md`. The pipeline alternates between rev... |
pg_090__seed_42 | pg_090 | 42 | [
"litigation-manager",
"legal-strategist",
"shadow-counsel",
"discovery-specialist",
"evidentiary-auditor"
] | [
{
"id": "f1",
"text": "You operate on the plan the litigation-manager has assigned you to this round. Read `lawsuit.md`, your assigned plan's `PLAN.md`, and your assigned plan's `THEORY.md`, then execute the evidentiary discovery:\n\n- Include every relevant custodian. Create specific search strings and cus... | f10 | {
"litigation-manager": [
"f5"
],
"legal-strategist": [
"f11",
"f6",
"f2",
"f12",
"f9"
],
"shadow-counsel": [
"f11",
"f2",
"f4",
"f7",
"f9"
],
"discovery-specialist": [
"f11",
"f1",
"f8",
"f3",
"f9"
],
"evidentiary-auditor": [
"f11",
... | I need you to set up a 5-agent pipeline that manages corporate litigation defense by developing abstract legal argumentation strategy and concrete evidentiary discovery in parallel, with each side informing the other. A lawsuit arrives as a natural-language statement in `lawsuit.md`. The pipeline alternates between rev... | I need you to set up a 5-agent pipeline that manages corporate litigation defense by developing abstract legal argumentation strategy and concrete evidentiary discovery in parallel, with each side informing the other. A lawsuit arrives as a natural-language statement in `lawsuit.md`. The pipeline alternates between rev... |
pg_091__seed_1 | pg_091 | 1 | [
"program-director",
"pharmacologist",
"toxicology-skeptic",
"trial-designer",
"regulatory-auditor"
] | [
{
"id": "f1",
"text": "Keep `SOLUTION.md` lean. Its body — everything except the `<review>` block — must stay under 3000 words (~20KB), about a four-page protocol summary; record only the final trial design and move iteration-by-iteration history into a separate `CHANGELOG.md`. Bold text must not exceed 5% ... | f10 | {
"program-director": [
"f2"
],
"pharmacologist": [
"f3",
"f5",
"f7",
"f11",
"f6"
],
"toxicology-skeptic": [
"f3",
"f7",
"f12",
"f4",
"f6"
],
"trial-designer": [
"f3",
"f8",
"f1",
"f13",
"f6"
],
"regulatory-auditor": [
"f3",
"f9",... | I need you to set up a 5-agent pipeline that develops pharmaceutical drugs by pairing theoretical biochemical mechanism-of-action models with the design of concrete clinical trial protocols in parallel, with each side informing the other. A disease target and therapeutic goals arrive as a natural-language statement in ... | I need you to set up a 5-agent pipeline that develops pharmaceutical drugs by pairing theoretical biochemical mechanism-of-action models with the design of concrete clinical trial protocols in parallel, with each side informing the other. A disease target and therapeutic goals arrive as a natural-language statement in ... |
pg_091__seed_42 | pg_091 | 42 | [
"program-director",
"pharmacologist",
"toxicology-skeptic",
"trial-designer",
"regulatory-auditor"
] | [
{
"id": "f1",
"text": "You operate on the plan the program-director has assigned you to this round. Read `problem.md`, your assigned plan's `PLAN.md`, and your assigned plan's `THEORY.md`, then write the concrete clinical trial protocol:\n\n- Outline every phase element. Define the patient inclusion/exclusi... | f10 | {
"program-director": [
"f5"
],
"pharmacologist": [
"f11",
"f6",
"f2",
"f12",
"f9"
],
"toxicology-skeptic": [
"f11",
"f2",
"f4",
"f7",
"f9"
],
"trial-designer": [
"f11",
"f1",
"f8",
"f3",
"f9"
],
"regulatory-auditor": [
"f11",
"f1... | I need you to set up a 5-agent pipeline that develops pharmaceutical drugs by pairing theoretical biochemical mechanism-of-action models with the design of concrete clinical trial protocols in parallel, with each side informing the other. A disease target and therapeutic goals arrive as a natural-language statement in ... | I need you to set up a 5-agent pipeline that develops pharmaceutical drugs by pairing theoretical biochemical mechanism-of-action models with the design of concrete clinical trial protocols in parallel, with each side informing the other. A disease target and therapeutic goals arrive as a natural-language statement in ... |
pg_092__seed_1 | pg_092 | 1 | [
"mission-controller",
"intelligence-analyst",
"red-team-adversary",
"field-tactician",
"security-auditor"
] | [
{
"id": "f1",
"text": "Keep `TACTICS.md` lean. Its body — everything except the `<review>` block — must stay under 3000 words (~20KB), about a four-page brief; record only the final operational state and move iteration-by-iteration planning history into a separate `CHANGELOG.md`. Bold text must not exceed 5... | f10 | {
"mission-controller": [
"f2"
],
"intelligence-analyst": [
"f3",
"f5",
"f7",
"f11",
"f6"
],
"red-team-adversary": [
"f3",
"f7",
"f12",
"f4",
"f6"
],
"field-tactician": [
"f3",
"f8",
"f1",
"f13",
"f6"
],
"security-auditor": [
"f3",
... | I need you to set up a 5-agent pipeline that develops covert intelligence operations by planning high-level geopolitical threat assessments and ground-level field tactics in parallel, with each side informing the other. A mission arrives as a natural-language directive in `mission_brief.md`. The pipeline alternates bet... | I need you to set up a 5-agent pipeline that develops covert intelligence operations by planning high-level geopolitical threat assessments and ground-level field tactics in parallel, with each side informing the other. A mission arrives as a natural-language directive in `mission_brief.md`. The pipeline alternates bet... |
pg_092__seed_42 | pg_092 | 42 | [
"mission-controller",
"intelligence-analyst",
"red-team-adversary",
"field-tactician",
"security-auditor"
] | [
{
"id": "f1",
"text": "You operate on the plan the mission-controller has assigned you to this round. Read `mission_brief.md`, your assigned plan's `PLAN.md`, and your assigned plan's `ASSESSMENT.md`, then structure the logistical plan:\n\n- Detail every asset, team composition, equipment loadout, and timel... | f10 | {
"mission-controller": [
"f5"
],
"intelligence-analyst": [
"f11",
"f6",
"f2",
"f12",
"f9"
],
"red-team-adversary": [
"f11",
"f2",
"f4",
"f7",
"f9"
],
"field-tactician": [
"f11",
"f1",
"f8",
"f3",
"f9"
],
"security-auditor": [
"f11",
... | I need you to set up a 5-agent pipeline that develops covert intelligence operations by planning high-level geopolitical threat assessments and ground-level field tactics in parallel, with each side informing the other. A mission arrives as a natural-language directive in `mission_brief.md`. The pipeline alternates bet... | I need you to set up a 5-agent pipeline that develops covert intelligence operations by planning high-level geopolitical threat assessments and ground-level field tactics in parallel, with each side informing the other. A mission arrives as a natural-language directive in `mission_brief.md`. The pipeline alternates bet... |
pg_093__seed_1 | pg_093 | 1 | [
"showrunner",
"narrative-architect",
"plot-hole-critic",
"scene-writer",
"continuity-auditor"
] | [
{
"id": "f1",
"text": "Keep `SCRIPT.md` focused. Its body — everything except the `<review>` block — must stay under 3000 words (~20KB), about a 12-page script sequence; record only the current draft state and move earlier draft history into a separate `CHANGELOG.md`. Bold text must not exceed 5% of the fil... | f10 | {
"showrunner": [
"f2"
],
"narrative-architect": [
"f3",
"f5",
"f7",
"f11",
"f6"
],
"plot-hole-critic": [
"f3",
"f7",
"f12",
"f4",
"f6"
],
"scene-writer": [
"f3",
"f8",
"f1",
"f13",
"f6"
],
"continuity-auditor": [
"f3",
"f9",
... | I need you to set up a 5-agent pipeline that develops feature film scripts by alternating between high-level narrative structure outlining and concrete scene-by-scene dialogue generation, with each side informing the other. A story concept arrives as a natural-language statement in `premise.md`. The pipeline alternates... | I need you to set up a 5-agent pipeline that develops feature film scripts by alternating between high-level narrative structure outlining and concrete scene-by-scene dialogue generation, with each side informing the other. A story concept arrives as a natural-language statement in `premise.md`. The pipeline alternates... |
pg_093__seed_42 | pg_093 | 42 | [
"showrunner",
"narrative-architect",
"plot-hole-critic",
"scene-writer",
"continuity-auditor"
] | [
{
"id": "f1",
"text": "You operate on the plan the showrunner has assigned you to this round. Read `premise.md`, your assigned plan's `PLAN.md`, and your assigned plan's `OUTLINE.md`, then write the screenplay pages:\n\n- Include standard formatting. Create proper sluglines (scene headings), action descript... | f10 | {
"showrunner": [
"f5"
],
"narrative-architect": [
"f11",
"f6",
"f2",
"f12",
"f9"
],
"plot-hole-critic": [
"f11",
"f2",
"f4",
"f7",
"f9"
],
"scene-writer": [
"f11",
"f1",
"f8",
"f3",
"f9"
],
"continuity-auditor": [
"f11",
"f13",
... | I need you to set up a 5-agent pipeline that develops feature film scripts by alternating between high-level narrative structure outlining and concrete scene-by-scene dialogue generation, with each side informing the other. A story concept arrives as a natural-language statement in `premise.md`. The pipeline alternates... | I need you to set up a 5-agent pipeline that develops feature film scripts by alternating between high-level narrative structure outlining and concrete scene-by-scene dialogue generation, with each side informing the other. A story concept arrives as a natural-language statement in `premise.md`. The pipeline alternates... |
pg_094__seed_1 | pg_094 | 1 | [
"campaign-manager",
"policy-architect",
"opposition-researcher",
"mobilization-director",
"compliance-auditor"
] | [
{
"id": "f1",
"text": "Keep `MOBILIZATION.md` lean. Its body — everything except the `<review>` block — must stay under 3000 words (~20KB), about a four-page memo; record only the final state and move iteration-by-iteration history into a separate `CHANGELOG.md`. Bold text must not exceed 5% of the file: if... | f10 | {
"campaign-manager": [
"f2"
],
"policy-architect": [
"f3",
"f5",
"f7",
"f11",
"f6"
],
"opposition-researcher": [
"f3",
"f7",
"f12",
"f4",
"f6"
],
"mobilization-director": [
"f3",
"f8",
"f1",
"f13",
"f6"
],
"compliance-auditor": [
"f3... | I need you to set up a 5-agent pipeline that orchestrates a political campaign by developing central policy platforms and localized ground-game voter mobilization in parallel, with each side informing the other. A campaign challenge arrives as a natural-language statement in `problem.md`. The pipeline alternates betwee... | I need you to set up a 5-agent pipeline that orchestrates a political campaign by developing central policy platforms and localized ground-game voter mobilization in parallel, with each side informing the other. A campaign challenge arrives as a natural-language statement in `problem.md`. The pipeline alternates betwee... |
pg_094__seed_42 | pg_094 | 42 | [
"campaign-manager",
"policy-architect",
"opposition-researcher",
"mobilization-director",
"compliance-auditor"
] | [
{
"id": "f1",
"text": "You operate on the plan the campaign-manager has assigned you to this round. Read `problem.md`, your assigned plan's `PLAN.md`, and your assigned plan's `POLICY.md`, then implement the ground-game targeting and resource allocation scripts:\n\n- Include every import. Create each variab... | f10 | {
"campaign-manager": [
"f5"
],
"policy-architect": [
"f11",
"f6",
"f2",
"f12",
"f9"
],
"opposition-researcher": [
"f11",
"f2",
"f4",
"f7",
"f9"
],
"mobilization-director": [
"f11",
"f1",
"f8",
"f3",
"f9"
],
"compliance-auditor": [
"f... | I need you to set up a 5-agent pipeline that orchestrates a political campaign by developing central policy platforms and localized ground-game voter mobilization in parallel, with each side informing the other. A campaign challenge arrives as a natural-language statement in `problem.md`. The pipeline alternates betwee... | I need you to set up a 5-agent pipeline that orchestrates a political campaign by developing central policy platforms and localized ground-game voter mobilization in parallel, with each side informing the other. A campaign challenge arrives as a natural-language statement in `problem.md`. The pipeline alternates betwee... |
pg_095__seed_1 | pg_095 | 1 | [
"care coordinator",
"diagnostician",
"diagnostic challenge physician",
"treatment planner",
"patient-safety reviewer"
] | [
{
"id": "f1",
"text": "Keep `MANAGEMENT.md` lean. Its body — everything except the `<review>` block — must stay under 3000 words (~20KB), about a four-page note; record only the final state and move iteration-by-iteration history into a separate `CHANGELOG.md`. Bold text must not exceed 5% of the file: if e... | f10 | {
"care coordinator": [
"f2"
],
"diagnostician": [
"f3",
"f5",
"f7",
"f11",
"f6"
],
"diagnostic challenge physician": [
"f3",
"f7",
"f12",
"f4",
"f6"
],
"treatment planner": [
"f3",
"f8",
"f1",
"f13",
"f6"
],
"patient-safety reviewer": [
... | I need you to set up a 5-agent pipeline that manages a complex patient case by developing diagnostic reasoning and treatment planning in parallel, with each side informing the other. A patient case arrives as a natural-language clinical summary in `case.md`. The pipeline alternates between revising the diagnostic asses... | I need you to set up a 5-agent pipeline that manages a complex patient case by developing diagnostic reasoning and treatment planning in parallel, with each side informing the other. A patient case arrives as a natural-language clinical summary in `case.md`. The pipeline alternates between revising the diagnostic asses... |
pg_095__seed_42 | pg_095 | 42 | [
"care coordinator",
"diagnostician",
"diagnostic challenge physician",
"treatment planner",
"patient-safety reviewer"
] | [
{
"id": "f1",
"text": "You operate on the plan the care coordinator has assigned you to this round. Read `case.md`, your assigned plan's `PLAN.md`, and your assigned plan's `DIAGNOSIS.md`, then draft the management plan:\n\n- Include every proposed order or action needed for the plan: medications with dose/... | f10 | {
"care coordinator": [
"f5"
],
"diagnostician": [
"f11",
"f6",
"f2",
"f12",
"f9"
],
"diagnostic challenge physician": [
"f11",
"f2",
"f4",
"f7",
"f9"
],
"treatment planner": [
"f11",
"f1",
"f8",
"f3",
"f9"
],
"patient-safety reviewer": [... | I need you to set up a 5-agent pipeline that manages a complex patient case by developing diagnostic reasoning and treatment planning in parallel, with each side informing the other. A patient case arrives as a natural-language clinical summary in `case.md`. The pipeline alternates between revising the diagnostic asses... | I need you to set up a 5-agent pipeline that manages a complex patient case by developing diagnostic reasoning and treatment planning in parallel, with each side informing the other. A patient case arrives as a natural-language clinical summary in `case.md`. The pipeline alternates between revising the diagnostic asses... |
pg_096__seed_1 | pg_096 | 1 | [
"litigation coordinator",
"defense strategist",
"mock prosecutor",
"case investigator",
"evidence challenge attorney"
] | [
{
"id": "f1",
"text": "Keep `EVIDENCE_WORKUP.md` lean. Its body — everything except the `<review>` block — must stay under 3000 words (~20KB), about a four-page note; record only the current final state and move iteration-by-iteration history into a separate `CHANGELOG.md`. Bold text must not exceed 5% of t... | f10 | {
"litigation coordinator": [
"f2"
],
"defense strategist": [
"f3",
"f5",
"f7",
"f11",
"f6"
],
"mock prosecutor": [
"f3",
"f7",
"f12",
"f4",
"f6"
],
"case investigator": [
"f3",
"f8",
"f1",
"f13",
"f6"
],
"evidence challenge attorney": [
... | I need you to set up a 5-agent pipeline for criminal trial preparation. The team develops a defense theory of the case and an evidence/witness workup in parallel, with each side informing the other. A case arrives as a neutral case packet in `case.md`: charges, alleged facts, discovery summaries, known witnesses, key d... | I need you to set up a 5-agent pipeline for criminal trial preparation. The team develops a defense theory of the case and an evidence/witness workup in parallel, with each side informing the other. A case arrives as a neutral case packet in `case.md`: charges, alleged facts, discovery summaries, known witnesses, key d... |
pg_096__seed_42 | pg_096 | 42 | [
"litigation coordinator",
"defense strategist",
"mock prosecutor",
"case investigator",
"evidence challenge attorney"
] | [
{
"id": "f1",
"text": "You operate on the plan the litigation coordinator has assigned you to this round. Read `case.md`, your assigned plan's `PLAN.md`, and your assigned plan's `DEFENSE_THEORY.md`, then build the evidence and witness workup:\n\n- Identify every witness, exhibit, recording, report, physica... | f10 | {
"litigation coordinator": [
"f5"
],
"defense strategist": [
"f11",
"f6",
"f2",
"f12",
"f9"
],
"mock prosecutor": [
"f11",
"f2",
"f4",
"f7",
"f9"
],
"case investigator": [
"f11",
"f1",
"f8",
"f3",
"f9"
],
"evidence challenge attorney": [... | I need you to set up a 5-agent pipeline for criminal trial preparation. The team develops a defense theory of the case and an evidence/witness workup in parallel, with each side informing the other. A case arrives as a neutral case packet in `case.md`: charges, alleged facts, discovery summaries, known witnesses, key d... | I need you to set up a 5-agent pipeline for criminal trial preparation. The team develops a defense theory of the case and an evidence/witness workup in parallel, with each side informing the other. A case arrives as a neutral case packet in `case.md`: charges, alleged facts, discovery summaries, known witnesses, key d... |
pg_097__seed_1 | pg_097 | 1 | [
"collections coordinator",
"provenance researcher",
"authentication skeptic",
"conservation scientist",
"technical examination auditor"
] | [
{
"id": "f1",
"text": "Keep `TESTING.md` lean. Its body — everything except the `<review>` block — must stay under 3000 words (~20KB), about a four-page examination note; record only the final state and move iteration-by-iteration history into a separate `CHANGELOG.md`. Bold text must not exceed 5% of the f... | f10 | {
"collections coordinator": [
"f2"
],
"provenance researcher": [
"f3",
"f5",
"f7",
"f11",
"f6"
],
"authentication skeptic": [
"f3",
"f7",
"f12",
"f4",
"f6"
],
"conservation scientist": [
"f3",
"f8",
"f1",
"f13",
"f6"
],
"technical examin... | I need you to set up a 5-agent pipeline that evaluates a museum object's authenticity by developing provenance and attribution arguments in parallel with material-testing documentation, with each side informing the other. An object arrives as a curatorial brief in `artifact.md`. The pipeline alternates between revising... | I need you to set up a 5-agent pipeline that evaluates a museum object's authenticity by developing provenance and attribution arguments in parallel with material-testing documentation, with each side informing the other. An object arrives as a curatorial brief in `artifact.md`. The pipeline alternates between revising... |
pg_097__seed_42 | pg_097 | 42 | [
"collections coordinator",
"provenance researcher",
"authentication skeptic",
"conservation scientist",
"technical examination auditor"
] | [
{
"id": "f1",
"text": "You operate on the plan the collections coordinator has assigned you to this round. Read `artifact.md`, your assigned plan's `PLAN.md`, and your assigned plan's `PROVENANCE.md`, then produce the technical examination and testing record:\n\n- Include every method used. Record instrumen... | f10 | {
"collections coordinator": [
"f5"
],
"provenance researcher": [
"f11",
"f6",
"f2",
"f12",
"f9"
],
"authentication skeptic": [
"f11",
"f2",
"f4",
"f7",
"f9"
],
"conservation scientist": [
"f11",
"f1",
"f8",
"f3",
"f9"
],
"technical exami... | I need you to set up a 5-agent pipeline that evaluates a museum object's authenticity by developing provenance and attribution arguments in parallel with material-testing documentation, with each side informing the other. An object arrives as a curatorial brief in `artifact.md`. The pipeline alternates between revising... | I need you to set up a 5-agent pipeline that evaluates a museum object's authenticity by developing provenance and attribution arguments in parallel with material-testing documentation, with each side informing the other. An object arrives as a curatorial brief in `artifact.md`. The pipeline alternates between revising... |
pg_098__seed_1 | pg_098 | 1 | [
"mission operations coordinator",
"anomaly analyst",
"fault-review board member",
"recovery procedure engineer",
"flight-safety verifier"
] | [
{
"id": "f1",
"text": "Keep `RECOVERY_PROCEDURE.md` lean. Its body — everything except the `<review>` block — must stay under 3000 words (~20KB), about a four-page note; record only the final state and move iteration-by-iteration history into a separate `CHANGELOG.md`. Bold text must not exceed 5% of the fi... | f10 | {
"mission operations coordinator": [
"f2"
],
"anomaly analyst": [
"f3",
"f5",
"f7",
"f11",
"f6"
],
"fault-review board member": [
"f3",
"f7",
"f12",
"f4",
"f6"
],
"recovery procedure engineer": [
"f3",
"f8",
"f1",
"f13",
"f6"
],
"flight-... | I need you to set up a 5-agent pipeline that responds to spacecraft anomalies by developing fault-cause analysis and recovery procedures in parallel, with each side informing the other under independent mission-safety review. An anomaly arrives as a natural-language incident report in `anomaly.md`. The pipeline alterna... | I need you to set up a 5-agent pipeline that responds to spacecraft anomalies by developing fault-cause analysis and recovery procedures in parallel, with each side informing the other under independent mission-safety review. An anomaly arrives as a natural-language incident report in `anomaly.md`. The pipeline alterna... |
pg_098__seed_42 | pg_098 | 42 | [
"mission operations coordinator",
"anomaly analyst",
"fault-review board member",
"recovery procedure engineer",
"flight-safety verifier"
] | [
{
"id": "f1",
"text": "You operate on the plan the mission operations coordinator has assigned you to this round. Read `anomaly.md`, your assigned plan's `PLAN.md`, and your assigned plan's `CAUSE_ANALYSIS.md`, then implement the recovery procedure:\n\n- Include every command, prerequisite, parameter value,... | f10 | {
"mission operations coordinator": [
"f5"
],
"anomaly analyst": [
"f11",
"f6",
"f2",
"f12",
"f9"
],
"fault-review board member": [
"f11",
"f2",
"f4",
"f7",
"f9"
],
"recovery procedure engineer": [
"f11",
"f1",
"f8",
"f3",
"f9"
],
"flight... | I need you to set up a 5-agent pipeline that responds to spacecraft anomalies by developing fault-cause analysis and recovery procedures in parallel, with each side informing the other under independent mission-safety review. An anomaly arrives as a natural-language incident report in `anomaly.md`. The pipeline alterna... | I need you to set up a 5-agent pipeline that responds to spacecraft anomalies by developing fault-cause analysis and recovery procedures in parallel, with each side informing the other under independent mission-safety review. An anomaly arrives as a natural-language incident report in `anomaly.md`. The pipeline alterna... |
pg_099__seed_1 | pg_099 | 1 | [
"permit coordinator",
"impact assessor",
"agency scientific reviewer",
"mitigation planner",
"compliance inspector"
] | [
{
"id": "f1",
"text": "Keep `MITIGATION.md` lean. Its body — everything except the `<review>` block — must stay under 3000 words (~20KB), about a four-page note; record only the final state and move iteration-by-iteration history into a separate `CHANGELOG.md`. Bold text must not exceed 5% of the file: if e... | f10 | {
"permit coordinator": [
"f2"
],
"impact assessor": [
"f3",
"f5",
"f7",
"f11",
"f6"
],
"agency scientific reviewer": [
"f3",
"f7",
"f12",
"f4",
"f6"
],
"mitigation planner": [
"f3",
"f8",
"f1",
"f13",
"f6"
],
"compliance inspector": [
... | I need you to set up a 5-agent pipeline that supports environmental impact permitting for a proposed infrastructure project by developing ecological impact claims and mitigation plans in parallel, with each side informing the other. A project arrives as a natural-language description in `project.md`. The pipeline alter... | I need you to set up a 5-agent pipeline that supports environmental impact permitting for a proposed infrastructure project by developing ecological impact claims and mitigation plans in parallel, with each side informing the other. A project arrives as a natural-language description in `project.md`. The pipeline alter... |
pg_099__seed_42 | pg_099 | 42 | [
"permit coordinator",
"impact assessor",
"agency scientific reviewer",
"mitigation planner",
"compliance inspector"
] | [
{
"id": "f1",
"text": "You operate on the plan the permit coordinator has assigned you to this round. Read `project.md`, your assigned plan's `PLAN.md`, and your assigned plan's `IMPACT.md`, then prepare the mitigation package:\n\n- Include every required mitigation component. Identify each impact the plan ... | f10 | {
"permit coordinator": [
"f5"
],
"impact assessor": [
"f11",
"f6",
"f2",
"f12",
"f9"
],
"agency scientific reviewer": [
"f11",
"f2",
"f4",
"f7",
"f9"
],
"mitigation planner": [
"f11",
"f1",
"f8",
"f3",
"f9"
],
"compliance inspector": [
... | I need you to set up a 5-agent pipeline that supports environmental impact permitting for a proposed infrastructure project by developing ecological impact claims and mitigation plans in parallel, with each side informing the other. A project arrives as a natural-language description in `project.md`. The pipeline alter... | I need you to set up a 5-agent pipeline that supports environmental impact permitting for a proposed infrastructure project by developing ecological impact claims and mitigation plans in parallel, with each side informing the other. A project arrives as a natural-language description in `project.md`. The pipeline alter... |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.