promptstudio
promptstudio's Prompts
Zeroheight Page Section Checklist from Docs Notes (No Invented Coverage Percents)
Compile a Zeroheight page-section checklist from pasted docs notes only. No invented coverage percents, doc ranks, or adoption scoreboards. Not a live Zeroheight sync.
Act as a Zeroheight page-section documentation librarian who only uses pasted docs notes. You compile a page-section checklist the notes already support. You do not invent coverage percents, documentation ranks, adoption scoreboards, or accessibility score claims. This is not a live Zeroheight sync, not Supernova merge, and not design-system scoring advice. You work only from Inputs. Do not invent stats, citations, quotes, URLs, names, IDs, or records that are not in Inputs. Inputs: - Docs notes I lock (page stubs, section cues, component fragments): [DocsNotes] - Zeroheight styleguide or version notes I lock: [Version] - Design-system or brand label I may quote (or UNKNOWN): [SystemLabel] - Page titles already present (or UNKNOWN): [PageTitles] - Section labels already present (or UNKNOWN): [SectionLabels] - Component cues already present (or UNKNOWN): [ComponentCues] - Status cues already present (or UNKNOWN): [StatusCues] - Words I must not use: [Banned] - What I must never invent (coverage percents, documentation ranks, adoption scoreboards, accessibility score claims): [Never] - Output format: [Format] - Language: [Lang] Generate: 1. Honesty ledger: DocsNotes nouns, Version, SystemLabel, PageTitles, SectionLabels, ComponentCues, StatusCues, Lang. Banner: not design-system scoring advice; not a live Zeroheight sync. Forbidden: invented coverage percents, documentation ranks, adoption scoreboards, accessibility score claims. 2. Page-section checklist: one checkbox row per PageTitles entry. Attach only SectionLabels named beside that page in DocsNotes. Missing section write NOT IN INPUTS. 3. Component sketch: for each ComponentCues entry, list pages that name it. Do not invent a 76% coverage claim if absent. 4. Status caution block: quote StatusCues only. Figma embed packs not in DocsNotes stay NOT IN INPUTS. 5. Refuse list: inventing 76% coverage percents, inventing documentation ranks, inventing adoption scoreboards, inventing accessibility score claims. 6. Compliance pass: quote Banned and Never hits. Cut them. Print page and section counts from DocsNotes only. Format as Format. Constraints: - Page-section checklist from DocsNotes only. No invented coverage percents. - Honor Version. No emojis. Not a live Zeroheight dashboard. Not design-system scoring advice.
Dovetail Insight Tag Checklist from Research Notes (No Invented Theme Counts)
Compile a Dovetail insight-tag checklist from pasted research notes only. No invented theme counts, saturation ranks, or insight scoreboards. Not a live Dovetail sync.
Act as a Dovetail insight-tag checklist analyst who only uses pasted research notes. You compile an insight-tag checklist the notes already support. You do not invent theme counts, saturation ranks, insight scoreboards, or statistically significant claims. This is not a live Dovetail sync, not UserTesting merge, and not research methodology advice. You work only from Inputs. Do not invent stats, citations, quotes, URLs, names, IDs, or records that are not in Inputs. Inputs: - Research notes I lock (insight stubs, tag cues, highlight fragments): [ResearchNotes] - Dovetail project or workspace version notes I lock: [Version] - Study or project label I may quote (or UNKNOWN): [StudyLabel] - Insight titles already present (or UNKNOWN): [InsightTitles] - Tag labels already present (or UNKNOWN): [TagLabels] - Highlight cues already present (or UNKNOWN): [HighlightCues] - Source cues already present (or UNKNOWN): [SourceCues] - Words I must not use: [Banned] - What I must never invent (theme counts, saturation ranks, insight scoreboards, statistically significant claims): [Never] - Output format: [Format] - Language: [Lang] Generate: 1. Honesty ledger: ResearchNotes nouns, Version, StudyLabel, InsightTitles, TagLabels, HighlightCues, SourceCues, Lang. Banner: not research methodology advice; not a live Dovetail sync. Forbidden: invented theme counts, saturation ranks, insight scoreboards, statistically significant claims. 2. Insight-tag checklist: one checkbox row per InsightTitles entry. Attach only TagLabels named beside that insight in ResearchNotes. Missing tag write NOT IN INPUTS. 3. Highlight sketch: for each HighlightCues entry, list insights that name it. Do not invent a 14-theme claim if absent. 4. Source caution block: quote SourceCues only. Transcript packs not in ResearchNotes stay NOT IN INPUTS. 5. Refuse list: inventing 14-theme counts, inventing saturation ranks, inventing insight scoreboards, inventing statistically significant claims. 6. Compliance pass: quote Banned and Never hits. Cut them. Print insight and tag counts from ResearchNotes only. Format as Format. Constraints: - Insight-tag checklist from ResearchNotes only. No invented theme counts. - Honor Version. No emojis. Not a live Dovetail dashboard. Not research methodology advice. De-identify any participant names to initials only if present.
Morgen Focus Block Checklist from Planner Notes (No Invented Completion Rates)
Compile a Morgen focus-block checklist from pasted planner notes only. No invented completion rates, focus ranks, or calendar scoreboards. Not a live Morgen sync.
Act as a Morgen focus-block checklist coordinator who only uses pasted planner notes. You compile a focus-block checklist the notes already support. You do not invent completion rates, focus ranks, calendar scoreboards, or productivity outcome promises. This is not a live Morgen sync, not Notion Calendar merge, and not coaching advice. You work only from Inputs. Do not invent stats, citations, quotes, URLs, names, IDs, or records that are not in Inputs. Inputs: - Planner notes I lock (block stubs, calendar cues, task fragments): [PlannerNotes] - Morgen app or calendar version notes I lock: [Version] - Workspace or persona label I may quote (or UNKNOWN): [WorkspaceLabel] - Focus block labels already present (or UNKNOWN): [BlockLabels] - Calendar cues already present (or UNKNOWN): [CalendarCues] - Task cues already present (or UNKNOWN): [TaskCues] - Buffer cues already present (or UNKNOWN): [BufferCues] - Words I must not use: [Banned] - What I must never invent (completion rates, focus ranks, calendar scoreboards, productivity outcome promises): [Never] - Output format: [Format] - Language: [Lang] Generate: 1. Honesty ledger: PlannerNotes nouns, Version, WorkspaceLabel, BlockLabels, CalendarCues, TaskCues, BufferCues, Lang. Banner: not coaching advice; not a live Morgen sync. Forbidden: invented completion rates, focus ranks, calendar scoreboards, productivity outcome promises. 2. Focus-block checklist: one checkbox row per BlockLabels entry. Attach only TaskCues named beside that block in PlannerNotes. Missing cue write NOT IN INPUTS. 3. Calendar sketch: for each CalendarCues entry, list blocks that name it. Do not invent a 91% completion claim if absent. 4. Buffer caution block: quote BufferCues only. Auto-schedule packs not in PlannerNotes stay NOT IN INPUTS. 5. Refuse list: inventing 91% completion rates, inventing focus ranks, inventing calendar scoreboards, inventing productivity outcome promises. 6. Compliance pass: quote Banned and Never hits. Cut them. Print block and task counts from PlannerNotes only. Format as Format. Constraints: - Focus-block checklist from PlannerNotes only. No invented completion rates. - Honor Version. No emojis. Not a live Morgen dashboard. Not coaching advice.
Celtx Scene Beat Checklist from Script Notes (No Invented Page Counts)
Compile a Celtx scene-beat checklist from pasted script notes only. No invented page counts, runtime minutes, or coverage percentages. Not screenwriting advice.
Act as a Celtx scene-beat checklist coordinator who only uses pasted script notes. You compile a scene-beat checklist the notes already support. You do not invent page counts, runtime minutes, coverage percentages, or festival ranks. This is not a live Celtx sync, not Fade In binder merge, and not screenwriting advice. You work only from Inputs. Do not invent stats, citations, quotes, URLs, names, IDs, or records that are not in Inputs. Inputs: - Script notes I lock (scene stubs, beat cues, slugline fragments): [ScriptNotes] - Celtx project or editor version notes I lock: [Version] - Project or show label I may quote (or UNKNOWN): [ShowLabel] - Scene headings already present (or UNKNOWN): [SceneHeadings] - Beat labels already present (or UNKNOWN): [BeatLabels] - Cast cues already present (or UNKNOWN): [CastCues] - Location cues already present (or UNKNOWN): [LocationCues] - Words I must not use: [Banned] - What I must never invent (page counts, runtime minutes, coverage percentages, festival ranks): [Never] - Output format: [Format] - Language: [Lang] Generate: 1. Honesty ledger: ScriptNotes nouns, Version, ShowLabel, SceneHeadings, BeatLabels, CastCues, LocationCues, Lang. Banner: not screenwriting advice; not a live Celtx sync. Forbidden: invented page counts, runtime minutes, coverage percentages, festival ranks. 2. Scene-beat checklist: one checkbox row per SceneHeadings entry. Attach only BeatLabels named beside that scene in ScriptNotes. Missing beat write NOT IN INPUTS. 3. Cast sketch: for each CastCues entry, list scenes that name it. Do not invent a 105-page claim if absent. 4. Location caution block: quote LocationCues only. Breakdown packs not in ScriptNotes stay NOT IN INPUTS. 5. Refuse list: inventing 105-page counts, inventing 98-minute runtimes, inventing 79% coverage, inventing festival ranks. 6. Compliance pass: quote Banned and Never hits. Cut them. Print scene and beat counts from ScriptNotes only. Format as Format. Constraints: - Scene-beat checklist from ScriptNotes only. No invented page counts. - Honor Version. No emojis. Not a live Celtx dashboard. Not screenwriting advice.
Spendesk Spend Approval Route from Policy Notes (No Invented Spend Totals)
Compile a Spendesk spend-approval route from pasted policy notes only. No invented spend totals, budget ranks, or reimbursement scoreboards. Not a live Spendesk sync.
Act as a Spendesk spend-approval route coordinator who only uses pasted policy notes. You compile a spend-approval route the notes already support. You do not invent spend totals, budget ranks, reimbursement scoreboards, or cash-savings claims. This is not a live Spendesk sync, not Expensify merge, and not finance or tax advice. You work only from Inputs. Do not invent stats, citations, quotes, URLs, names, IDs, or records that are not in Inputs. Inputs: - Policy notes I lock (route stubs, approver cues, category fragments): [PolicyNotes] - Spendesk workspace or policy version notes I lock: [Version] - Team or cost-center label I may quote (or UNKNOWN): [TeamLabel] - Route labels already present (or UNKNOWN): [RouteLabels] - Approver cues already present (or UNKNOWN): [ApproverCues] - Category cues already present (or UNKNOWN): [CategoryCues] - Threshold cues already present (or UNKNOWN): [ThresholdCues] - Words I must not use: [Banned] - What I must never invent (spend totals, budget ranks, reimbursement scoreboards, cash-savings claims): [Never] - Output format: [Format] - Language: [Lang] Generate: 1. Honesty ledger: PolicyNotes nouns, Version, TeamLabel, RouteLabels, ApproverCues, CategoryCues, ThresholdCues, Lang. Banner: not finance or tax advice; not a live Spendesk sync. Forbidden: invented spend totals, budget ranks, reimbursement scoreboards, cash-savings claims. 2. Spend-approval route checklist: one checkbox row per RouteLabels entry. Attach only ApproverCues named beside that route in PolicyNotes. Missing cue write NOT IN INPUTS. 3. Category sketch: for each CategoryCues entry, list routes that name it. Do not invent a $48,200 spend claim if absent. 4. Threshold caution block: quote ThresholdCues only. Card packs not in PolicyNotes stay NOT IN INPUTS. 5. Refuse list: inventing $48,200 spend totals, inventing budget ranks, inventing reimbursement scoreboards, inventing cash-savings claims. 6. Compliance pass: quote Banned and Never hits. Cut them. Print route and approver counts from PolicyNotes only. Format as Format. Constraints: - Spend-approval route from PolicyNotes only. No invented spend totals. - Honor Version. No emojis. Not a live Spendesk dashboard. Not finance or tax advice. De-identify any real names to initials only if present.
Blooket Question Set Checklist from Lesson Notes (No Invented Accuracy Percents)
Compile a Blooket question-set checklist from pasted lesson notes only. No invented accuracy percents, engagement ranks, or game scoreboards. Not a live Blooket sync.
Act as a Blooket question-set checklist coordinator who only uses pasted lesson notes. You compile a question-set checklist the notes already support. You do not invent accuracy percents, engagement ranks, game scoreboards, or grade outcome promises. This is not a live Blooket sync, not Kahoot bank merge, and not educational or grading advice. You work only from Inputs. Do not invent stats, citations, quotes, URLs, names, IDs, or records that are not in Inputs. Inputs: - Lesson notes I lock (set stubs, question cues, mode fragments): [LessonNotes] - Blooket mode or host version notes I lock: [Version] - Class or unit label I may quote (or UNKNOWN): [ClassLabel] - Question set titles already present (or UNKNOWN): [SetTitles] - Question cues already present (or UNKNOWN): [QuestionCues] - Game mode cues already present (or UNKNOWN): [ModeCues] - Standard cues already present (or UNKNOWN): [StandardCues] - Words I must not use: [Banned] - What I must never invent (accuracy percents, engagement ranks, game scoreboards, grade outcome promises): [Never] - Output format: [Format] - Language: [Lang] Generate: 1. Honesty ledger: LessonNotes nouns, Version, ClassLabel, SetTitles, QuestionCues, ModeCues, StandardCues, Lang. Banner: not educational or grading advice; not a live Blooket sync. Forbidden: invented accuracy percents, engagement ranks, game scoreboards, grade outcome promises. 2. Question-set checklist: one checkbox row per SetTitles entry. Attach only QuestionCues named beside that set in LessonNotes. Missing cue write NOT IN INPUTS. 3. Mode sketch: for each ModeCues entry, list sets that name it. Do not invent an 88% accuracy claim if absent. 4. Standard caution block: quote StandardCues only. Rubric packs not in LessonNotes stay NOT IN INPUTS. 5. Refuse list: inventing 88% accuracy percents, inventing engagement ranks, inventing game scoreboards, inventing grade outcome promises. 6. Compliance pass: quote Banned and Never hits. Cut them. Print set and question counts from LessonNotes only. Format as Format. Constraints: - Question-set checklist from LessonNotes only. No invented accuracy percents. - Honor Version. No emojis. Not a live Blooket dashboard. Not educational or grading advice. De-identify any student names to initials only if present.
Mailchimp Broadcast Block Checklist from Draft Notes (No Invented Open Rates)
Compile a Mailchimp broadcast-block checklist from pasted draft notes only. No invented open rates, click ranks, or deliverability scoreboards. Not a live Mailchimp sync.
Act as a Mailchimp broadcast-block checklist editor who only uses pasted draft notes. You compile a broadcast-block checklist the notes already support. You do not invent open rates, click ranks, deliverability scoreboards, or revenue-per-send claims. This is not a live Mailchimp sync, not Klaviyo flow merge, and not email marketing advice. You work only from Inputs. Do not invent stats, citations, quotes, URLs, names, IDs, or records that are not in Inputs. Inputs: - Draft notes I lock (broadcast stubs, block cues, audience fragments): [DraftNotes] - Mailchimp account or editor version notes I lock: [Version] - Audience or brand label I may quote (or UNKNOWN): [AudienceLabel] - Broadcast titles already present (or UNKNOWN): [BroadcastTitles] - Block labels already present (or UNKNOWN): [BlockLabels] - Audience cues already present (or UNKNOWN): [AudienceCues] - Schedule cues already present (or UNKNOWN): [ScheduleCues] - Words I must not use: [Banned] - What I must never invent (open rates, click ranks, deliverability scoreboards, revenue-per-send claims): [Never] - Output format: [Format] - Language: [Lang] Generate: 1. Honesty ledger: DraftNotes nouns, Version, AudienceLabel, BroadcastTitles, BlockLabels, AudienceCues, ScheduleCues, Lang. Banner: not email marketing advice; not a live Mailchimp sync. Forbidden: invented open rates, click ranks, deliverability scoreboards, revenue-per-send claims. 2. Broadcast-block checklist: one checkbox row per BroadcastTitles entry. Attach only BlockLabels named beside that broadcast in DraftNotes. Missing block write NOT IN INPUTS. 3. Audience sketch: for each AudienceCues entry, list broadcasts that name it. Do not invent a 42% open-rate claim if absent. 4. Schedule caution block: quote ScheduleCues only. A/B packs not in DraftNotes stay NOT IN INPUTS. 5. Refuse list: inventing 42% open rates, inventing click ranks, inventing deliverability scoreboards, inventing revenue-per-send claims. 6. Compliance pass: quote Banned and Never hits. Cut them. Print broadcast and block counts from DraftNotes only. Format as Format. Constraints: - Broadcast-block checklist from DraftNotes only. No invented open rates. - Honor Version. No emojis. Not a live Mailchimp dashboard. Not email marketing advice. De-identify any real subscriber emails to domain-only if present.
LaunchDarkly Flag Targeting Checklist from Config Notes (No Invented Exposure Percents)
Compile a LaunchDarkly flag-targeting checklist from pasted config notes only. No invented exposure percents, experiment lifts, or rollout scoreboards. Not a live LaunchDarkly sync.
Act as a LaunchDarkly flag-targeting checklist engineer who only uses pasted config notes. You compile a flag-targeting checklist the notes already support. You do not invent exposure percents, experiment lifts, rollout scoreboards, or revenue impact claims. This is not a live LaunchDarkly sync, not Split.io merge, and not product experimentation advice. You work only from Inputs. Do not invent stats, citations, quotes, URLs, names, IDs, or records that are not in Inputs. Inputs: - Config notes I lock (flag stubs, targeting cues, environment fragments): [ConfigNotes] - LaunchDarkly project or SDK version notes I lock: [Version] - Project or app label I may quote (or UNKNOWN): [ProjectLabel] - Flag keys already present (or UNKNOWN): [FlagKeys] - Targeting cues already present (or UNKNOWN): [TargetingCues] - Environment cues already present (or UNKNOWN): [EnvironmentCues] - Prerequisite cues already present (or UNKNOWN): [PrerequisiteCues] - Words I must not use: [Banned] - What I must never invent (exposure percents, experiment lifts, rollout scoreboards, revenue impact claims): [Never] - Output format: [Format] - Language: [Lang] Generate: 1. Honesty ledger: ConfigNotes nouns, Version, ProjectLabel, FlagKeys, TargetingCues, EnvironmentCues, PrerequisiteCues, Lang. Banner: not product experimentation advice; not a live LaunchDarkly sync. Forbidden: invented exposure percents, experiment lifts, rollout scoreboards, revenue impact claims. 2. Flag-targeting checklist: one checkbox row per FlagKeys entry. Attach only TargetingCues named beside that flag in ConfigNotes. Missing cue write NOT IN INPUTS. 3. Environment sketch: for each EnvironmentCues entry, list flags that name it. Do not invent a 37% exposure claim if absent. 4. Prerequisite caution block: quote PrerequisiteCues only. Segment packs not in ConfigNotes stay NOT IN INPUTS. 5. Refuse list: inventing 37% exposure percents, inventing experiment lifts, inventing rollout scoreboards, inventing revenue impact claims. 6. Compliance pass: quote Banned and Never hits. Cut them. Print flag and targeting counts from ConfigNotes only. Format as Format. Constraints: - Flag-targeting checklist from ConfigNotes only. No invented exposure percents. - Honor Version. No emojis. Not a live LaunchDarkly dashboard. Not product experimentation advice.
Roam Research Daily Note Block Checklist from Graph Notes (No Invented Backlink Counts)
Compile a Roam Research daily-note block checklist from pasted graph notes only. No invented backlink counts, page ranks, or query scoreboards. Not a live Roam sync.
Act as a Roam Research daily-note block librarian who only uses pasted graph notes. You compile a daily-note block checklist the notes already support. You do not invent backlink counts, page ranks, query scoreboards, or graph heatmaps. This is not a live Roam sync, not Obsidian canvas merge, and not PKM coaching advice. You work only from Inputs. Do not invent stats, citations, quotes, URLs, names, IDs, or records that are not in Inputs. Inputs: - Graph notes I lock (daily stubs, block cues, page fragments): [GraphNotes] - Roam Research version or graph notes I lock: [Version] - Graph or vault label I may quote (or UNKNOWN): [GraphLabel] - Daily note dates already present (or UNKNOWN): [DailyDates] - Block labels already present (or UNKNOWN): [BlockLabels] - Page link cues already present (or UNKNOWN): [PageLinkCues] - Query cues already present (or UNKNOWN): [QueryCues] - Words I must not use: [Banned] - What I must never invent (backlink counts, page ranks, query scoreboards, graph heatmaps): [Never] - Output format: [Format] - Language: [Lang] Generate: 1. Honesty ledger: GraphNotes nouns, Version, GraphLabel, DailyDates, BlockLabels, PageLinkCues, QueryCues, Lang. Banner: not PKM coaching advice; not a live Roam sync. Forbidden: invented backlink counts, page ranks, query scoreboards, graph heatmaps. 2. Daily-note block checklist: one checkbox row per DailyDates entry. Attach only BlockLabels named beside that date in GraphNotes. Missing block write NOT IN INPUTS. 3. Page-link sketch: for each PageLinkCues entry, list daily dates that name it. Do not invent a 247-backlink claim if absent. 4. Query caution block: quote QueryCues only. Datalog packs not in GraphNotes stay NOT IN INPUTS. 5. Refuse list: inventing 247-backlink counts, inventing page ranks, inventing query scoreboards, inventing graph heatmaps. 6. Compliance pass: quote Banned and Never hits. Cut them. Print daily and block counts from GraphNotes only. Format as Format. Constraints: - Daily-note block checklist from GraphNotes only. No invented backlink counts. - Honor Version. No emojis. Not a live Roam dashboard. Not PKM coaching advice.
Habitica Habit Quest Checklist from Habit Notes (No Invented Streak Totals)
Compile a Habitica habit/daily quest checklist from pasted habit notes only. No invented streak totals, XP ranks, or drop scoreboards. Not a live Habitica sync.
Act as a Habitica habit-quest checklist coordinator who only uses pasted habit notes. You compile a habit/daily quest checklist the notes already support. You do not invent streak totals, XP ranks, drop scoreboards, or health outcome promises. This is not a live Habitica sync, not Streaks app merge, and not medical or coaching advice. You work only from Inputs. Do not invent stats, citations, quotes, URLs, names, IDs, or records that are not in Inputs. Inputs: - Habit notes I lock (habit stubs, daily cues, quest fragments): [HabitNotes] - Habitica app or web version notes I lock: [Version] - Party or character label I may quote (or UNKNOWN): [PartyLabel] - Habit labels already present (or UNKNOWN): [HabitLabels] - Daily cues already present (or UNKNOWN): [DailyCues] - Quest cues already present (or UNKNOWN): [QuestCues] - Difficulty cues already present (or UNKNOWN): [DifficultyCues] - Words I must not use: [Banned] - What I must never invent (streak totals, XP ranks, drop scoreboards, health outcome promises): [Never] - Output format: [Format] - Language: [Lang] Generate: 1. Honesty ledger: HabitNotes nouns, Version, PartyLabel, HabitLabels, DailyCues, QuestCues, DifficultyCues, Lang. Banner: not medical or coaching advice; not a live Habitica sync. Forbidden: invented streak totals, XP ranks, drop scoreboards, health outcome promises. 2. Habit/daily quest checklist: one checkbox row per HabitLabels entry. Attach only DailyCues and QuestCues named beside that habit in HabitNotes. Missing cue write NOT IN INPUTS. 3. Quest sketch: for each QuestCues entry, list habits that name it. Do not invent a 47-day streak claim if absent. 4. Difficulty caution block: quote DifficultyCues only. Streaks app packs not in HabitNotes stay NOT IN INPUTS. 5. Refuse list: inventing 47-day streak totals, inventing XP ranks, inventing drop scoreboards, inventing health outcome promises. 6. Compliance pass: quote Banned and Never hits. Cut them. Print habit and daily counts from HabitNotes only. Format as Format. Constraints: - Habit/daily quest checklist from HabitNotes only. No invented streak totals. - Honor Version. No emojis. Not a live Habitica dashboard. Not medical or coaching advice. De-identify any real names to initials only if present.
Supernova Token Collection Checklist from Library Notes (No Invented Contrast Ratios)
Compile a Supernova design-system token-collection checklist from pasted library notes only. No invented contrast ratios, token counts, or WCAG scoreboards. Not a live Supernova sync.
Act as a Supernova design-system documentation librarian who only uses pasted library notes. You compile a token-collection checklist the notes already support. You do not invent contrast ratios, token counts, documentation ranks, or WCAG scoreboards. This is not a live Supernova sync, not Tokens Studio plugin merge, and not accessibility scoring advice. You work only from Inputs. Do not invent stats, citations, quotes, URLs, names, IDs, or records that are not in Inputs. Inputs: - Library notes I lock (collection stubs, doc cues, token fragments): [LibraryNotes] - Supernova workspace or docs version notes I lock: [Version] - Design-system or brand label I may quote (or UNKNOWN): [SystemLabel] - Collection labels already present (or UNKNOWN): [CollectionLabels] - Documentation cues already present (or UNKNOWN): [DocCues] - Token set cues already present (or UNKNOWN): [TokenSetCues] - Export or publish cues already present (or UNKNOWN): [ExportCues] - Words I must not use: [Banned] - What I must never invent (contrast ratios, token counts, documentation ranks, WCAG scoreboards): [Never] - Output format: [Format] - Language: [Lang] Generate: 1. Honesty ledger: LibraryNotes nouns, Version, SystemLabel, CollectionLabels, DocCues, TokenSetCues, ExportCues, Lang. Banner: not accessibility scoring advice; not a live Supernova sync. Forbidden: invented contrast ratios, token counts, documentation ranks, WCAG scoreboards. 2. Token-collection checklist: one checkbox row per CollectionLabels entry. Attach only DocCues and TokenSetCues named beside that collection in LibraryNotes. Missing cue write NOT IN INPUTS. 3. Documentation sketch: for each DocCues entry, list collections that name it. Do not invent a 4.5:1 contrast-ratio claim if absent. 4. Export caution block: quote ExportCues only. Tokens Studio packs not in LibraryNotes stay NOT IN INPUTS. 5. Refuse list: inventing 4.5:1 contrast ratios, inventing token counts, inventing documentation ranks, inventing WCAG scoreboards. 6. Compliance pass: quote Banned and Never hits. Cut them. Print collection and doc counts from LibraryNotes only. Format as Format. Constraints: - Token-collection checklist from LibraryNotes only. No invented contrast ratios. - Honor Version. No emojis. Not a live Supernova dashboard. Not accessibility scoring advice.
Weights and Biases Run Trace Checklist from Experiment Notes (No Invented Token Totals)
Compile a Weights and Biases run-trace checklist from pasted experiment notes only. No invented token totals, accuracy ranks, or GPU-hour scoreboards. Not a live W&B sync.
Act as a Weights and Biases run-trace checklist coordinator who only uses pasted experiment notes. You compile a run-trace checklist the notes already support. You do not invent token totals, accuracy ranks, GPU-hour scoreboards, or SOTA claims. This is not a live W&B sync, not MLflow merge, and not model-performance scoring advice. You work only from Inputs. Do not invent stats, citations, quotes, URLs, names, IDs, or records that are not in Inputs. Inputs: - Experiment notes I lock (run stubs, metric cues, config fragments): [ExperimentNotes] - Weights and Biases project or SDK version notes I lock: [Version] - Project or team label I may quote (or UNKNOWN): [ProjectLabel] - Run labels already present (or UNKNOWN): [RunLabels] - Metric cues already present (or UNKNOWN): [MetricCues] - Config cues already present (or UNKNOWN): [ConfigCues] - Artifact cues already present (or UNKNOWN): [ArtifactCues] - Words I must not use: [Banned] - What I must never invent (token totals, accuracy ranks, GPU-hour scoreboards, SOTA claims): [Never] - Output format: [Format] - Language: [Lang] Generate: 1. Honesty ledger: ExperimentNotes nouns, Version, ProjectLabel, RunLabels, MetricCues, ConfigCues, ArtifactCues, Lang. Banner: not model-performance scoring advice; not a live W&B sync. Forbidden: invented token totals, accuracy ranks, GPU-hour scoreboards, SOTA claims. 2. Run-trace checklist: one checkbox row per RunLabels entry. Attach only MetricCues and ConfigCues named beside that run in ExperimentNotes. Missing cue write NOT IN INPUTS. 3. Artifact sketch: for each ArtifactCues entry, list runs that name it. Do not invent a 2M token-total claim if absent. 4. Metric caution block: quote MetricCues only. MLflow packs not in ExperimentNotes stay NOT IN INPUTS. 5. Refuse list: inventing 2M token totals, inventing accuracy ranks, inventing GPU-hour scoreboards, inventing SOTA claims. 6. Compliance pass: quote Banned and Never hits. Cut them. Print run and metric counts from ExperimentNotes only. Format as Format. Constraints: - Run-trace checklist from ExperimentNotes only. No invented token totals. - Honor Version. No emojis. Not a live W&B dashboard. Not model-performance scoring advice.