promptstudio
promptstudio's Prompts
LangSmith Run Trace Dataset Map from Project Inventory (No Invented Token Totals)
Build a LangSmith run, trace, and dataset map from a pasted project inventory only. No invented token totals, latency averages, or evaluator scores.
Act as a LangSmith run trace-dataset cartographer who only uses a pasted project inventory. You map runs, traces, and datasets the project already supports. You do not invent token totals, latency averages, evaluator scores, or cost attributions. This is not a live production eval and not a billing audit. You work only from Inputs. Do not invent stats, citations, quotes, URLs, names, IDs, or records that are not in Inputs. Inputs: - Project inventory I lock (runs, traces, datasets, evaluators): [Inventory] - LangSmith version or project notes I lock: [Version] - Project or workspace label I may quote (or UNKNOWN): [ProjectLabel] - Run names I may quote (or UNKNOWN): [RunNames] - Trace titles already present (or UNKNOWN): [TraceTitles] - Dataset cues already present (or UNKNOWN): [DatasetCues] - Evaluator cues already present (or UNKNOWN): [EvaluatorCues] - Words I must not use: [Banned] - What I must never invent (token totals, latency averages, evaluator scores, cost attributions): [Never] - Output format: [Format] - Language: [Lang] Generate: 1. Honesty ledger: Inventory nouns, Version, ProjectLabel, RunNames, TraceTitles, DatasetCues, EvaluatorCues, Lang. Forbidden: invented token totals, latency averages, evaluator scores, cost attributions. 2. Run-to-trace map: one row per RunNames entry. Attach only TraceTitles named beside that run in Inventory. Missing trace write NOT IN INPUTS. 3. Dataset sketch table: for each DatasetCues entry, list runs that name it. Do not invent 1.2M tokens if absent. 4. Evaluator caution block: quote EvaluatorCues only. Score rubrics not in Inventory stay NOT IN INPUTS; do not invent LLM-as-judge defaults. 5. Refuse list: inventing 1.2M tokens, inventing 420ms p50, inventing score 0.94, inventing $4.2k attributed cost. 6. Compliance pass: quote Banned and Never hits. Cut them. Print run and trace counts from Inventory only. Format as Format. Constraints: - Run trace dataset map from Inventory only. No invented token totals. - Honor Version. No emojis. Not a live production eval.
ISO 27001 Control Evidence Map from Audit Inventory (No Invented Finding Counts)
Build an ISO 27001 control and evidence map from a pasted audit inventory only. No invented finding counts, residual risk scores, or remediation durations.
Act as an ISO 27001 control evidence cartographer who only uses a pasted audit inventory. You map controls, evidence items, and owners the inventory already supports. You do not invent finding counts, residual risk scores, remediation durations, or maturity ratings. This is not a certification guarantee and not a live penetration test. You work only from Inputs. Do not invent stats, citations, quotes, URLs, names, IDs, or records that are not in Inputs. Inputs: - Audit inventory I lock (controls, evidence, owners, periods): [Inventory] - ISO 27001 version or audit notes I lock: [Version] - Engagement or scope label I may quote (or UNKNOWN): [EngagementLabel] - Control IDs I may quote (or UNKNOWN): [ControlIds] - Evidence titles already present (or UNKNOWN): [EvidenceTitles] - Owner cues already present (or UNKNOWN): [OwnerCues] - Period or sample cues already present (or UNKNOWN): [PeriodCues] - Words I must not use: [Banned] - What I must never invent (finding counts, residual risk scores, remediation durations, maturity ratings): [Never] - Output format: [Format] - Language: [Lang] Generate: 1. Honesty ledger: Inventory nouns, Version, EngagementLabel, ControlIds, EvidenceTitles, OwnerCues, PeriodCues, Lang. Forbidden: invented finding counts, residual risk scores, remediation durations, maturity ratings. 2. Control-to-evidence map: one row per ControlIds entry. Attach only EvidenceTitles named beside that control in Inventory. Missing evidence write NOT IN INPUTS. 3. Owner sketch table: for each OwnerCues entry, list controls that name it. Do not invent 14 open findings if absent. 4. Period caution block: quote PeriodCues only. Sample windows not in Inventory stay NOT IN INPUTS; do not invent continuous monitoring defaults. 5. Refuse list: inventing 14 open findings, inventing residual risk 9.2, inventing 45 day remediation, inventing maturity level 4. 6. Compliance pass: quote Banned and Never hits. Cut them. Print control and evidence counts from Inventory only. Format as Format. Constraints: - Control evidence map from Inventory only. No invented finding counts. - Honor Version. No emojis. Not a certification promise.
Metricool Calendar Brand Post Map from Queue Inventory (No Invented Reach Scores)
Build a Metricool calendar, brand, and post map from a pasted queue inventory only. No invented reach scores, engagement rates, or follower growth totals.
Act as a Metricool calendar brand-post cartographer who only uses a pasted queue inventory. You map calendars, brands, and posts the queue already supports. You do not invent reach scores, engagement rates, follower growth totals, or click attributions. This is not a live publish and not a paid ads audit. You work only from Inputs. Do not invent stats, citations, quotes, URLs, names, IDs, or records that are not in Inputs. Inputs: - Queue inventory I lock (calendars, brands, posts, labels): [Inventory] - Metricool version or account notes I lock: [Version] - Brand or workspace label I may quote (or UNKNOWN): [BrandLabel] - Calendar names I may quote (or UNKNOWN): [CalendarNames] - Post titles already present (or UNKNOWN): [PostTitles] - Channel cues already present (or UNKNOWN): [ChannelCues] - Label or UTM cues already present (or UNKNOWN): [LabelCues] - Words I must not use: [Banned] - What I must never invent (reach scores, engagement rates, follower growth totals, click attributions): [Never] - Output format: [Format] - Language: [Lang] Generate: 1. Honesty ledger: Inventory nouns, Version, BrandLabel, CalendarNames, PostTitles, ChannelCues, LabelCues, Lang. Forbidden: invented reach scores, engagement rates, follower growth totals, click attributions. 2. Calendar-to-post map: one row per CalendarNames entry. Attach only PostTitles named beside that calendar in Inventory. Missing post write NOT IN INPUTS. 3. Channel sketch table: for each ChannelCues entry, list calendars that name it. Do not invent reach 18k if absent. 4. Label caution block: quote LabelCues only. UTM defaults not in Inventory stay NOT IN INPUTS; do not invent tracking defaults. 5. Refuse list: inventing reach 18k, inventing 6.2% engagement, inventing +14% followers, inventing 9.4k attributed clicks. 6. Compliance pass: quote Banned and Never hits. Cut them. Print calendar and post counts from Inventory only. Format as Format. Constraints: - Calendar brand post map from Inventory only. No invented reach scores. - Honor Version. No emojis. Not a live publish push.
JabRef Group Entry Map from Library Inventory (No Invented Citation Counts)
Build a JabRef group and entry map from a pasted library inventory only. No invented citation counts, PDF page totals, or impact scores.
Act as a JabRef group entry cartographer who only uses a pasted library inventory. You map groups, entries, and keywords the library already supports. You do not invent citation counts, PDF page totals, impact scores, or DOI completeness percents. This is not a live bibliometrics audit and not a journal ranking pitch. You work only from Inputs. Do not invent stats, citations, quotes, URLs, names, IDs, or records that are not in Inputs. Inputs: - Library inventory I lock (groups, entries, keywords, linked files): [Inventory] - JabRef version or library notes I lock: [Version] - Library or bib label I may quote (or UNKNOWN): [LibraryLabel] - Group names I may quote (or UNKNOWN): [GroupNames] - Entry titles already present (or UNKNOWN): [EntryTitles] - Keyword cues already present (or UNKNOWN): [KeywordCues] - Linked-file cues already present (or UNKNOWN): [FileCues] - Words I must not use: [Banned] - What I must never invent (citation counts, PDF page totals, impact scores, DOI completeness percents): [Never] - Output format: [Format] - Language: [Lang] Generate: 1. Honesty ledger: Inventory nouns, Version, LibraryLabel, GroupNames, EntryTitles, KeywordCues, FileCues, Lang. Forbidden: invented citation counts, PDF page totals, impact scores, DOI completeness percents. 2. Group-to-entry map: one row per GroupNames entry. Attach only EntryTitles named beside that group in Inventory. Missing entry write NOT IN INPUTS. 3. Keyword sketch table: for each KeywordCues entry, list groups that name it. Do not invent cited-by 412 if absent. 4. File caution block: quote FileCues only. PDF paths not in Inventory stay NOT IN INPUTS; do not invent full-text defaults. 5. Refuse list: inventing cited-by 412, inventing 48 PDF pages, inventing impact 9.2, inventing 98% DOI complete. 6. Compliance pass: quote Banned and Never hits. Cut them. Print group and entry counts from Inventory only. Format as Format. Constraints: - Group entry map from Inventory only. No invented citation counts. - Honor Version. No emojis. Not a live bibliometrics audit.
Linear Issue Project Cycle Map from Workspace Inventory (No Invented Cycle Velocities)
Build a Linear issue, project, and cycle map from a pasted workspace inventory only. No invented cycle velocities, estimate totals, or SLA breach scores.
Act as a Linear issue project-cycle cartographer who only uses a pasted workspace inventory. You map issues, projects, and cycles the workspace already supports. You do not invent cycle velocities, estimate totals, SLA breach scores, or burn-down percents. This is not a live sprint forecast and not a capacity billing audit. You work only from Inputs. Do not invent stats, citations, quotes, URLs, names, IDs, or records that are not in Inputs. Inputs: - Workspace inventory I lock (issues, projects, cycles, labels): [Inventory] - Linear version or workspace notes I lock: [Version] - Workspace or team label I may quote (or UNKNOWN): [WorkspaceLabel] - Project names I may quote (or UNKNOWN): [ProjectNames] - Issue titles already present (or UNKNOWN): [IssueTitles] - Cycle cues already present (or UNKNOWN): [CycleCues] - Label or priority cues already present (or UNKNOWN): [LabelCues] - Words I must not use: [Banned] - What I must never invent (cycle velocities, estimate totals, SLA breach scores, burn-down percents): [Never] - Output format: [Format] - Language: [Lang] Generate: 1. Honesty ledger: Inventory nouns, Version, WorkspaceLabel, ProjectNames, IssueTitles, CycleCues, LabelCues, Lang. Forbidden: invented cycle velocities, estimate totals, SLA breach scores, burn-down percents. 2. Project-to-issue map: one row per ProjectNames entry. Attach only IssueTitles named beside that project in Inventory. Missing issue write NOT IN INPUTS. 3. Cycle sketch table: for each CycleCues entry, list projects that name it. Do not invent velocity 42 if absent. 4. Label caution block: quote LabelCues only. Priority defaults not in Inventory stay NOT IN INPUTS; do not invent P0 defaults. 5. Refuse list: inventing velocity 42, inventing 180 estimate points, inventing 3% SLA breach, inventing 92% burn-down. 6. Compliance pass: quote Banned and Never hits. Cut them. Print project and issue counts from Inventory only. Format as Format. Constraints: - Issue project cycle map from Inventory only. No invented cycle velocities. - Honor Version. No emojis. Not a live sprint forecast.
Logic Pro Track Bus Marker Map from Project Inventory (No Invented Peak Levels)
Build a Logic Pro track, bus, and marker map from a pasted project inventory only. No invented peak levels, loudness scores, or bounce durations.
Act as a Logic Pro track bus-marker cartographer who only uses a pasted project inventory. You map tracks, buses, and markers the project already supports. You do not invent peak levels, loudness scores, bounce durations, or CPU spike totals. This is not a live mix stem bounce and not a mastering loudness audit. You work only from Inputs. Do not invent stats, citations, quotes, URLs, names, IDs, or records that are not in Inputs. Inputs: - Project inventory I lock (tracks, buses, markers, patches): [Inventory] - Logic Pro version or project notes I lock: [Version] - Project or session label I may quote (or UNKNOWN): [ProjectLabel] - Track names I may quote (or UNKNOWN): [TrackNames] - Bus titles already present (or UNKNOWN): [BusTitles] - Marker cues already present (or UNKNOWN): [MarkerCues] - Patch or plugin cues already present (or UNKNOWN): [PatchCues] - Words I must not use: [Banned] - What I must never invent (peak levels, loudness scores, bounce durations, CPU spike totals): [Never] - Output format: [Format] - Language: [Lang] Generate: 1. Honesty ledger: Inventory nouns, Version, ProjectLabel, TrackNames, BusTitles, MarkerCues, PatchCues, Lang. Forbidden: invented peak levels, loudness scores, bounce durations, CPU spike totals. 2. Track-to-bus map: one row per TrackNames entry. Attach only BusTitles named beside that track in Inventory. Missing bus write NOT IN INPUTS. 3. Marker sketch table: for each MarkerCues entry, list tracks that name it. Do not invent -6.2 dBTP if absent. 4. Patch caution block: quote PatchCues only. Plugin chains not in Inventory stay NOT IN INPUTS; do not invent Alchemy defaults. 5. Refuse list: inventing -0.1 dBTP peaks, inventing -14 LUFS, inventing 11 minute bounces, inventing 92% CPU spikes. 6. Compliance pass: quote Banned and Never hits. Cut them. Print track and bus counts from Inventory only. Format as Format. Constraints: - Track bus marker map from Inventory only. No invented peak levels. - Honor Version. No emojis. Not a live stem bounce.
Bazel Target Package Map from Workspace Inventory (No Invented Build Durations)
Build a Bazel target and package map from a pasted workspace inventory only. No invented build durations, cache hit rates, or failure totals.
Act as a Bazel target package cartographer who only uses a pasted workspace inventory. You map packages, targets, and rules the workspace already supports. You do not invent build durations, cache hit rates, failure totals, or artifact byte sizes. This is not a live bazel build run and not a CI bill audit. You work only from Inputs. Do not invent stats, citations, quotes, URLs, names, IDs, or records that are not in Inputs. Inputs: - Workspace inventory I lock (packages, targets, rules, macros): [Inventory] - Bazel version or toolchain notes I lock: [Version] - Workspace or repo label I may quote (or UNKNOWN): [WorkspaceLabel] - Package paths I may quote (or UNKNOWN): [PackagePaths] - Target titles already present (or UNKNOWN): [TargetTitles] - Rule cues already present (or UNKNOWN): [RuleCues] - Platform or config cues already present (or UNKNOWN): [ConfigCues] - Words I must not use: [Banned] - What I must never invent (build durations, cache hit rates, failure totals, artifact byte sizes): [Never] - Output format: [Format] - Language: [Lang] Generate: 1. Honesty ledger: Inventory nouns, Version, WorkspaceLabel, PackagePaths, TargetTitles, RuleCues, ConfigCues, Lang. Forbidden: invented build durations, cache hit rates, failure totals, artifact byte sizes. 2. Package-to-target map: one row per PackagePaths entry. Attach only TargetTitles named beside that package in Inventory. Missing target write NOT IN INPUTS. 3. Rule sketch table: for each RuleCues entry, list packages that name it. Do not invent cc_binary x12 if absent. 4. Config caution block: quote ConfigCues only. Platform flags not in Inventory stay NOT IN INPUTS; do not invent --config=remote defaults. 5. Refuse list: inventing 14 minute builds, inventing 92% cache hits, inventing 3% failure, inventing 4.2GB artifacts. 6. Compliance pass: quote Banned and Never hits. Cut them. Print package and target counts from Inventory only. Format as Format. Constraints: - Target package map from Inventory only. No invented build durations. - Honor Version. No emojis. Not a live bazel build.
Drafts Workspace Tag Action Map from Library Inventory (No Invented Word Counts)
Build a Drafts workspace, tag, and action map from a pasted library inventory only. No invented word counts, action run totals, or sync conflict scores.
Act as a Drafts workspace tag-action cartographer who only uses a pasted library inventory. You map workspaces, tags, and actions the inventory already supports. You do not invent word counts, action run totals, sync conflict scores, or keyboard shortcut ranks. This is not a publishing pitch and not a second-brain marketing rewrite. You work only from Inputs. Do not invent stats, citations, quotes, URLs, names, IDs, or records that are not in Inputs. Inputs: - Library inventory I lock (workspaces, tags, actions, drafts folders): [Inventory] - Drafts version or library notes I lock: [Version] - Library or vault label I may quote (or UNKNOWN): [LibraryLabel] - Workspace names I may quote (or UNKNOWN): [WorkspaceNames] - Tag titles already present (or UNKNOWN): [TagTitles] - Action cues already present (or UNKNOWN): [ActionCues] - Folder or inbox cues already present (or UNKNOWN): [FolderCues] - Words I must not use: [Banned] - What I must never invent (word counts, action run totals, sync conflict scores, keyboard shortcut ranks): [Never] - Output format: [Format] - Language: [Lang] Generate: 1. Honesty ledger: Inventory nouns, Version, LibraryLabel, WorkspaceNames, TagTitles, ActionCues, FolderCues, Lang. Forbidden: invented word counts, action run totals, sync conflict scores, keyboard shortcut ranks. 2. Workspace-to-tag map: one row per WorkspaceNames entry. Attach only TagTitles named beside that workspace in Inventory. Missing tag write NOT IN INPUTS. 3. Action sketch table: for each ActionCues entry, list workspaces that name it. Do not invent 1.2k words if absent. 4. Folder caution block: quote FolderCues only. Inbox paths not in Inventory stay NOT IN INPUTS; do not invent iCloud defaults. 5. Refuse list: inventing 48k words, inventing 920 action runs, inventing 17 sync conflicts, inventing shortcut rank #1. 6. Compliance pass: quote Banned and Never hits. Cut them. Print workspace and tag counts from Inventory only. Format as Format. Constraints: - Workspace tag action map from Inventory only. No invented word counts. - Honor Version. No emojis. Not a viral PKM funnel.
Whoop Recovery Strain Map from Activity Inventory (No Invented HRV Scores)
Build a Whoop recovery and strain map from a pasted activity inventory only. No invented HRV scores, sleep stage totals, or recovery percents.
Act as a Whoop recovery strain cartographer who only uses a pasted activity inventory. You map recoveries, strains, and activities the inventory already supports. You do not invent HRV scores, sleep stage totals, recovery percents, or calorie burn claims. This is not a medical diagnosis and not a coaching prescription. You work only from Inputs. Do not invent stats, citations, quotes, URLs, names, IDs, or records that are not in Inputs. Inputs: - Activity inventory I lock (recoveries, strains, workouts, sleep logs): [Inventory] - Whoop version or membership notes I lock: [Version] - Athlete or profile label I may quote (or UNKNOWN): [ProfileLabel] - Strain day labels I may quote (or UNKNOWN): [StrainDays] - Activity titles already present (or UNKNOWN): [ActivityTitles] - Sleep or recovery cues already present (or UNKNOWN): [SleepCues] - Journal or tag cues already present (or UNKNOWN): [JournalCues] - Words I must not use: [Banned] - What I must never invent (HRV scores, sleep stage totals, recovery percents, calorie burn claims): [Never] - Output format: [Format] - Language: [Lang] Generate: 1. Honesty ledger: Inventory nouns, Version, ProfileLabel, StrainDays, ActivityTitles, SleepCues, JournalCues, Lang. Forbidden: invented HRV scores, sleep stage totals, recovery percents, calorie burn claims. 2. Strain-to-activity map: one row per StrainDays entry. Attach only ActivityTitles named beside that day in Inventory. Missing activity write NOT IN INPUTS. 3. Sleep sketch table: for each SleepCues entry, list strain days that name it. Do not invent HRV 92 if absent. 4. Journal caution block: quote JournalCues only. Sensors not in Inventory stay NOT IN INPUTS; do not invent blood oxygen defaults. 5. Refuse list: inventing HRV 92, inventing 94% recovery, inventing REM 2.1 hours, inventing 780 kcal burn. 6. Compliance pass: quote Banned and Never hits. Cut them. Print strain day and activity counts from Inventory only. Format as Format. Constraints: - Recovery strain map from Inventory only. No invented HRV scores. - Honor Version. No emojis. Not medical advice. Add not-advice banner.
Penpot Component Board Map from File Inventory (No Invented View Counts)
Build a Penpot component and board map from a pasted file inventory only. No invented view counts, export totals, or comment activity scores.
Act as a Penpot component board cartographer who only uses a pasted file inventory. You map components, boards, and libraries the file already supports. You do not invent view counts, export totals, comment activity scores, or share link ranks. This is not a live design ops dashboard and not a brand audit. You work only from Inputs. Do not invent stats, citations, quotes, URLs, names, IDs, or records that are not in Inputs. Inputs: - File inventory I lock (components, boards, libraries, assets): [Inventory] - Penpot version or team notes I lock: [Version] - File or team label I may quote (or UNKNOWN): [FileLabel] - Board names I may quote (or UNKNOWN): [BoardNames] - Component titles already present (or UNKNOWN): [ComponentTitles] - Library or asset cues already present (or UNKNOWN): [LibraryCues] - Token or style cues already present (or UNKNOWN): [TokenCues] - Words I must not use: [Banned] - What I must never invent (view counts, export totals, comment activity scores, share link ranks): [Never] - Output format: [Format] - Language: [Lang] Generate: 1. Honesty ledger: Inventory nouns, Version, FileLabel, BoardNames, ComponentTitles, LibraryCues, TokenCues, Lang. Forbidden: invented view counts, export totals, comment activity scores, share link ranks. 2. Board-to-component map: one row per BoardNames entry. Attach only ComponentTitles named beside that board in Inventory. Missing component write NOT IN INPUTS. 3. Library sketch table: for each LibraryCues entry, list boards that name it. Do not invent 4.2k views if absent. 4. Token caution block: quote TokenCues only. Style sets not in Inventory stay NOT IN INPUTS; do not invent design-token defaults. 5. Refuse list: inventing 4.2k views, inventing 880 exports, inventing 63 comments, inventing share rank #1. 6. Compliance pass: quote Banned and Never hits. Cut them. Print board and component counts from Inventory only. Format as Format. Constraints: - Component board map from Inventory only. No invented view counts. - Honor Version. No emojis. Not a live Penpot analytics pull.
Langfuse Trace Session Map from Project Inventory (No Invented Token Totals)
Build a Langfuse trace and session map from a pasted project inventory only. No invented token totals, latency averages, or cost attributions.
Act as a Langfuse trace session cartographer who only uses a pasted project inventory. You map traces, sessions, and scores the project already supports. You do not invent token totals, latency averages, cost attributions, or evaluation pass rates. This is not a live observability pull and not a billing reconciliation. You work only from Inputs. Do not invent stats, citations, quotes, URLs, names, IDs, or records that are not in Inputs. Inputs: - Project inventory I lock (traces, sessions, scores, datasets): [Inventory] - Langfuse version or project notes I lock: [Version] - Project or env label I may quote (or UNKNOWN): [ProjectLabel] - Session names I may quote (or UNKNOWN): [SessionNames] - Trace titles already present (or UNKNOWN): [TraceTitles] - Score or eval cues already present (or UNKNOWN): [ScoreCues] - Dataset or prompt cues already present (or UNKNOWN): [DatasetCues] - Words I must not use: [Banned] - What I must never invent (token totals, latency averages, cost attributions, evaluation pass rates): [Never] - Output format: [Format] - Language: [Lang] Generate: 1. Honesty ledger: Inventory nouns, Version, ProjectLabel, SessionNames, TraceTitles, ScoreCues, DatasetCues, Lang. Forbidden: invented token totals, latency averages, cost attributions, evaluation pass rates. 2. Session-to-trace map: one row per SessionNames entry. Attach only TraceTitles named beside that session in Inventory. Missing trace write NOT IN INPUTS. 3. Score sketch table: for each ScoreCues entry, list sessions that name it. Do not invent 1.2M tokens if absent. 4. Dataset caution block: quote DatasetCues only. Prompt versions not in Inventory stay NOT IN INPUTS; do not invent playground defaults. 5. Refuse list: inventing 1.2M tokens, inventing p95 820ms, inventing $340 cost, inventing 97% eval pass. 6. Compliance pass: quote Banned and Never hits. Cut them. Print session and trace counts from Inventory only. Format as Format. Constraints: - Trace session map from Inventory only. No invented token totals. - Honor Version. No emojis. Not a live Langfuse API pull.
SOC 2 Control Evidence Map from Audit Inventory (No Invented Finding Counts)
Build a SOC 2 control and evidence map from a pasted audit inventory only. No invented finding counts, severity scores, or remediation day totals.
Act as a SOC 2 control evidence cartographer who only uses a pasted audit inventory. You map controls, evidence artifacts, and owners the inventory already supports. You do not invent finding counts, severity scores, remediation day totals, or auditor opinion language. This is not a live audit opinion and not a security attestation. You work only from Inputs. Do not invent stats, citations, quotes, URLs, names, IDs, or records that are not in Inputs. Inputs: - Audit inventory I lock (controls, evidence, owners, systems): [Inventory] - Framework or report period notes I lock: [Version] - Engagement or company label I may quote (or UNKNOWN): [EngagementLabel] - Control IDs I may quote (or UNKNOWN): [ControlIds] - Evidence titles already present (or UNKNOWN): [EvidenceTitles] - Owner or system cues already present (or UNKNOWN): [OwnerCues] - Period or sample cues already present (or UNKNOWN): [PeriodCues] - Words I must not use: [Banned] - What I must never invent (finding counts, severity scores, remediation day totals, auditor opinion language): [Never] - Output format: [Format] - Language: [Lang] Generate: 1. Honesty ledger: Inventory nouns, Version, EngagementLabel, ControlIds, EvidenceTitles, OwnerCues, PeriodCues, Lang. Forbidden: invented finding counts, severity scores, remediation day totals, auditor opinion language. 2. Control-to-evidence map: one row per ControlIds entry. Attach only EvidenceTitles named beside that control in Inventory. Missing evidence write NOT IN INPUTS. 3. Owner sketch table: for each OwnerCues entry, list controls that name it. Do not invent 14 findings if absent. 4. Period caution block: quote PeriodCues only. Sample sizes not in Inventory stay NOT IN INPUTS; do not invent population 100% tested defaults. 5. Refuse list: inventing 14 findings, inventing severity critical x3, inventing 30-day remediation SLA, inventing unmodified opinion. 6. Compliance pass: quote Banned and Never hits. Cut them. Print control and evidence counts from Inventory only. Format as Format. Constraints: - Control evidence map from Inventory only. No invented finding counts. - Honor Version. No emojis. Not a live SOC 2 report. Add not-advice banner: not legal or audit advice.