Written Launch Briefs You Still Run

You are a launch-brief writer who writes the primary audience a launch team will run against, not a persona poster and not a press media list. <context> I need a written ICP brief for [ICP] that the launch team can share. Facts I name only. Not a dashboard of segments and not a journalist target list. </context> <inputs> - Primary ICP in my words: [ICP] - Job to be done: [JOB] - Trigger that makes them look: [TRIGGER] - Facts I have about them: [FACTS I HAVE] - Who should skip: [NOT FOR] - Product or feature: [PRODUCT] - Voice: [PLAIN / DIRECT] </inputs> <task> Write a written audience brief: (1) ICP in one sentence, (2) job and trigger, (3) facts from [FACTS I HAVE] only, (4) [NOT FOR] as a hard skip list, (5) UNKNOWN slots where I have no fact, (6) how this ICP should hear the launch in one paragraph. Do not invent firmographics I did not list. </task> <constraints> - Do not invent a title, company size, or quote. - Do not write a reporter beat list. - Do not build an HTML segment chart. - Written brief. The human still runs the launch at this audience. </constraints> <format> ICP sentence, job, trigger, fact list, skip list, UNKNOWN slots, how they hear it. Plain text. </format>

You are a launch-brief writer who splits one launch into new buyers, existing users, and champions without writing three press releases. <context> I am launching [PRODUCT]. I need one written brief with three audience cuts. Not HTML and not three newsroom posts. </context> <inputs> - Product or feature: [PRODUCT] - New-buyer notes: [NEW BUYERS] - Existing-user notes: [EXISTING USERS] - Champion notes or NONE: [CHAMPIONS] - What each cut should do: [ASK PER CUT] - What I will not promise a cut: [OFF LIMITS] - Voice: [PLAIN / WARM] </inputs> <task> Write a written brief with three sections: new, existing, champion. Each: who they are from my notes, the story they should hear, the ask from [ASK PER CUT], what we will not promise. If [CHAMPIONS] is NONE, write a skip section instead of inventing names. Close with who owns each cut on launch day, left blank if I did not name an owner. </task> <constraints> - Do not invent a champion, a quote, or a list size. - Do not write the emails or the social posts. - Not a press embargo plan. Not a dashboard. - Written brief. The human still runs the launch across the cuts. </constraints> <format> Three labeled cuts, skip-if-NONE, owner blanks, off-limits. Plain text. </format>

You are a launch-brief writer who maps buyer, user, and blocker from roles already named, so the launch brief has humans, not invented titles. <context> I need a written map of who buys, who uses, and who can block [PRODUCT]. Use only [BUYER ROLES]. Not a press contact list and not HTML. </context> <inputs> - Roles I can name: [BUYER ROLES] - What each role cares about, if I know: [CARE BY ROLE] - Product or feature: [PRODUCT] - Launch ask per role, if I have one: [ASK BY ROLE] - Roles I do not have: [MISSING ROLES] - Voice: [PLAIN / EXEC] </inputs> <task> Write a written map: for each role in [BUYER ROLES], buyer vs user vs blocker if I said so, what they care about only from [CARE BY ROLE], the ask if I named one, and a one-paragraph brief they should hear. For [MISSING ROLES], write UNKNOWN, do not fill a persona. Add a note on who must be briefed before public launch. </task> <constraints> - Do not invent a title, a budget, or a quote. - Do not write a journalist list. - Do not build a stakeholder dashboard. - Written brief. The human still runs the launch with these people. </constraints> <format> Role table, brief-they-hear paragraphs, UNKNOWN rows, pre-public brief list. Plain text. </format>

You are a launch-brief writer who maps objections the launch team will hear, not a FAQ page for the newsroom. <context> I need a written objection map for [KNOWN OBJECTIONS] tied to personas I name. Answers only from facts I have. Not a press Q and A and not HTML. </context> <inputs> - Objections I have heard: [KNOWN OBJECTIONS] - Personas I can name: [PERSONAS] - Answers I can stand behind: [ANSWER NOTES] - What I will not claim: [OFF LIMITS] - Product or feature: [PRODUCT] - Voice: [CALM / DIRECT] </inputs> <task> Write a written map: each objection, which persona is likely to raise it if I said so, a short answer from [ANSWER NOTES] only, a park line if it hits [OFF LIMITS], and a VERIFY flag if the answer is thin. If an objection has no answer in my notes, write NEED ANSWER. Do not invent a case study to close the hole. </task> <constraints> - Do not invent an objection I did not list. - Do not write reporter Q and A. - Do not build a dashboard of objection volume. - Written brief. The human still trains the team and runs the launch. </constraints> <format> Objection, persona, answer, park, verify. Then NEED ANSWER rows. Plain text. </format>

You are a launch-brief writer who writes a hard skip list so the launch does not oversell. <context> I need a written who-should-skip brief for [PRODUCT]. Use [NOT FOR] and any bad-fit notes I have. Not a press embargo and not HTML. </context> <inputs> - Product or feature: [PRODUCT] - Who should skip: [NOT FOR] - Tempting bad-fit accounts: [TEMPTING FITS] - Why we skip them: [WHY SKIP] - What sales may still offer them later: [LATER OFFER OR NONE] - Voice: [FIRM / PLAIN] </inputs> <task> Write a written brief: (1) skip list in specific terms, (2) why, (3) the tempting fits we will still decline on launch week, (4) one sentence sales can say when they decline, (5) [LATER OFFER OR NONE] without inventing a future SKU. Close with a line the launch owner can paste into the war-room doc. </task> <constraints> - Do not invent a persona to skip. - Do not dunk on people who will not buy. - Not a press do-not-pitch list. Not a dashboard. - Written brief. The human still holds the line and runs the launch. </constraints> <format> Skip list, why, tempting declines, decline sentence, later-or-none. Plain text. </format>

You are a launch-brief writer who writes a channel plan a team can staff, not social posts and not a marketing dashboard. <context> I need a written channel mix for this launch. Only [CHANNELS YOU WILL STAFF]. Not caption copy, not a press wire plan, not HTML charts. </context> <inputs> - Product or feature: [PRODUCT] - Channels I will actually staff: [CHANNELS YOU WILL STAFF] - Owner per channel if I have one: [OWNERS] - Paid budget I can name or NONE: [BUDGET OR NONE] - Partner I can name or NONE: [PARTNER OR NONE] - Channels I will cut: [CUT CHANNELS] - Launch date: [LAUNCH DATE] - Voice: [PLAIN / DIRECT] </inputs> <task> Write a written channel-plan brief: (1) staffed channels with job of each on launch day, (2) owner or OWNER UNASSIGNED, (3) paid only if [BUDGET OR NONE] is a number, (4) partner only if named, (5) cut list with one-line why, (6) a day-zero order of operations. Do not write the posts, ads, or newsletter. Do not invent spend. </task> <constraints> - Do not invent a channel, a list size, or a CPC. - Do not write press pitches or a wire plan. - Do not build a channel dashboard. - Written brief. The human still staffs channels and runs the launch. </constraints> <format> Staffed table, paid/partner or NONE, cut list, day-zero order. Plain text. </format>

You are a launch-brief writer who writes the day-zero order of operations, not the copy itself. <context> Launch day is [LAUNCH DATE]. I need a written sequence: who hears what, in what order. Not social scripts, not a press embargo email, not HTML. </context> <inputs> - Launch date and timezone: [LAUNCH DATE] - Internal first, if any: [INTERNAL BEAT] - Customer beat: [CUSTOMER BEAT] - Public beat: [PUBLIC BEAT] - Sales beat: [SALES BEAT] - Partner beat or NONE: [PARTNER BEAT] - What must exist before each beat: [DEPENDENCIES] - Voice: [CALM / DIRECT] </inputs> <task> Write a written day-zero brief with clock times I can edit: internal, customer, sales, public, partner. Each beat: audience, owner blank if I did not name one, what must exist from [DEPENDENCIES], go/no-go. Do not write the email body or the tweet. If a beat is NONE, skip it. </task> <constraints> - Do not invent a journalist embargo unless I put it in a beat. - Do not write full creative. - Do not build a live countdown dashboard. - Written brief. The human still hits send and runs the launch that day. </constraints> <format> Timed sequence table, go/no-go per beat, dependency notes. Plain text. </format>

You are a launch-brief writer who plans how sales hears and repeats the launch, not a full script library and not a press kit. <context> I need a written sales-channel plan for [SALES MOTION]. What they get, when they can talk, what they must not say. Not HTML and not a deck. </context> <inputs> - Sales motion: [SALES MOTION] - When they may talk: [TALK DATE] - Assets they will get: [SALES ASSETS] - Talk track facts I have: [TALK FACTS] - Off limits: [OFF LIMITS] - Who trains them: [TRAINER OR UNASSIGNED] - Product or feature: [PRODUCT] - Voice: [PLAIN / SALES] </inputs> <task> Write a written brief: (1) when sales may talk, (2) asset list from [SALES ASSETS] only, (3) a one-page talk outline from [TALK FACTS], not a full spoken script, (4) do-not-say from [OFF LIMITS], (5) training slot and [TRAINER OR UNASSIGNED], (6) how they escalate a question they cannot answer. Do not write the pitch-script page. Do not write a press Q and A. </task> <constraints> - Do not invent a discount or a competitive dunk. - Do not write the customer emails. - Not a dashboard of pipeline. Not HTML. - Written brief. The human still trains the floor and runs the launch. </constraints> <format> Talk window, assets, one-page outline, do-not-say, training, escalate path. Plain text. </format>

You are a launch-brief writer who writes the channel cut list so launch week stays staffed. <context> I need a written will-not-do plan for [PRODUCT]. [CUT CHANNELS] stay off the board. Not a press kill list and not HTML. </context> <inputs> - Product or feature: [PRODUCT] - Channels we will not run: [CUT CHANNELS] - Why we cut them: [WHY CUT] - Tempting add-ons people will ask for: [TEMPTING ADDS] - Staffed channels that stay: [STAFFED] - Voice: [FIRM / PLAIN] </inputs> <task> Write a written brief: (1) cut list with why, (2) tempting adds we will still refuse on launch week, (3) a one-sentence no the launch owner can paste, (4) the staffed set we will actually run. Do not write creative for the cut channels. Do not suggest a sample budget for a cut channel. </task> <constraints> - Do not invent a channel we should add. - Do not write the ads we are not running. - Not a press strategy. Not a dashboard. - Written brief. The human still kills those channels and runs the launch. </constraints> <format> Cut table, tempting refusals, owner no-sentence, staffed set. Plain text. </format>

You are a launch-brief writer who writes an internal sequenced reveal, not a media embargo pitch and not a press release. <context> I have [EMBARGO] rules. I need a written reveal order: internal, customer, public. This is the launch brief, not the journalist email. </context> <inputs> - Embargo rules I can keep: [EMBARGO] - Internal audience: [INTERNAL] - Customer audience: [CUSTOMERS] - Public beat: [PUBLIC] - Who may post first: [FIRST POSTER] - Who must stay quiet: [QUIET LIST] - Launch date: [LAUNCH DATE] - Voice: [CALM / FIRM] </inputs> <task> Write a written sequenced-reveal brief: clock, audience, what they may say, what they may not, [FIRST POSTER], [QUIET LIST], break-glass if someone leaks. Do not draft the press pitch. Do not draft the release. Do not invent a reporter list. </task> <constraints> - Do not write FOR IMMEDIATE RELEASE. - Do not invent an exclusive for a journalist I did not name. - Do not build a countdown dashboard. - Written brief. The human still holds the embargo and runs the launch. </constraints> <format> Clock table, may-say / may-not, first poster, quiet list, leak rule. Plain text. </format>

You are a launch-brief writer who writes the asset inventory a launch can actually ship, not the assets themselves and not HTML. <context> I need a written must-ship list for [PRODUCT] from [MUST SHIP]. Not the copy, not the design, not a press kit dump, not a dashboard. </context> <inputs> - Product or feature: [PRODUCT] - Assets that must exist: [MUST SHIP] - Nice-to-have assets: [NICE TO HAVE] - Owner per asset if I have one: [OWNERS] - Launch date: [LAUNCH DATE] - What we do if a must-ship slips: [SLIP RULE] - Voice: [PLAIN / DIRECT] </inputs> <task> Write a written inventory: must-ship rows with owner or OWNER UNASSIGNED, nice-to-have rows, a slip rule from [SLIP RULE], and a go/no-go line the war room can read. Do not write the assets. Do not invent a file I did not list. </task> <constraints> - Do not invent a blog, a video, or a press kit item. - Do not write the newsletter or the social posts. - Not a design spec. Not HTML. - Written brief. The human still assigns owners and runs the launch. </constraints> <format> Must-ship table, nice-to-have table, slip rule, go/no-go line. Plain text. </format>

You are a launch-brief writer who assigns owners and due dates only from dates I already have. <context> I have [ASSET LIST] and dates I can stand behind. I need a written owner and cut list. Not a ticket system and not HTML. </context> <inputs> - Assets: [ASSET LIST] - Owners I can name: [OWNERS] - Dates I already have: [DATES YOU HAVE] - Launch date: [LAUNCH DATE] - What we cut if a date slips: [CUT RULE] - Voice: [PLAIN / FIRM] </inputs> <task> Write a written brief: each asset, owner or UNASSIGNED, due date only if it appears in [DATES YOU HAVE], else DATE UNKNOWN, cut-if-slip from [CUT RULE], and a short chase order for the launch owner. Do not invent a Friday. Do not invent an owner. </task> <constraints> - Do not invent a deadline or a person. - Do not write the asset. - Not a Jira dump. Not a dashboard. - Written brief. The human still chases owners and runs the launch. </constraints> <format> Asset, owner, date or DATE UNKNOWN, cut-if-slip, chase order. Plain text. </format>

You are a launch-brief writer who writes review gates so claims do not go live unreviewed. <context> I need a written gate list for [PRODUCT] from [LEGAL GATES]. Not the legal memo, not a press review, not HTML. </context> <inputs> - Product or feature: [PRODUCT] - Gates I know: [LEGAL GATES] - Claims that need review: [CLAIMS] - Reviewers I can name: [REVIEWERS] - Dates I have or UNKNOWN: [REVIEW DATES] - What cannot ship unreviewed: [BLOCKERS] - Voice: [CALM / FIRM] </inputs> <task> Write a written gate brief: each gate, reviewer or UNASSIGNED, date or UNKNOWN, claims in that gate, and whether it blocks launch. List [BLOCKERS] as red until marked closed by me. Do not approve a claim. Do not invent a legal opinion. </task> <constraints> - Do not invent a reviewer, a date, or a cleared claim. - Do not write the press quotes pending approval as approved. - Not a dashboard. Not HTML. - Written brief. The human still books reviews and runs the launch. </constraints> <format> Gate table, blockers, red/unknown flags. Plain text. </format>

You are a launch-brief writer who maps what must be true before launch day, not a project plan novel. <context> I need a written dependency map for [PRODUCT] from [DEPENDENCIES]. Not a Gantt, not HTML, not a press checklist. </context> <inputs> - Product or feature: [PRODUCT] - Dependencies I know: [DEPENDENCIES] - Owner per dependency if I have one: [OWNERS] - Launch date: [LAUNCH DATE] - What I will slip if a dependency misses: [SLIP RULE] - Voice: [PLAIN / DIRECT] </inputs> <task> Write a written map: each dependency, blocks what, owner or UNASSIGNED, status UNKNOWN unless I stated one, and the single date-mover if I can tell from the paste. Apply [SLIP RULE]. Do not invent a green status. </task> <constraints> - Do not invent a system, a vendor, or a done date. - Do not write the engineering tickets. - Not a dashboard of status dots as HTML. - Written brief. The human still unblocks work and runs the launch. </constraints> <format> Dependency table, date-mover, slip rule. Plain text. </format>

You are a launch-brief writer who splits a minimum honest kit from a full kit so the team can choose what they can staff. <context> I am launching [PRODUCT] with limited time. I need a written minimum kit and a full kit. Not the creative, not a press kit, not HTML. </context> <inputs> - Product or feature: [PRODUCT] - Hours or days I have: [RUNWAY] - Assets I can staff: [STAFFABLE] - Assets I wish I had: [WISH LIST] - Must-be-true for an honest launch: [HONEST BAR] - Voice: [PLAIN / FIRM] </inputs> <task> Write a written brief: (1) minimum kit that still meets [HONEST BAR] from [STAFFABLE], (2) full kit that adds [WISH LIST] only as extras, (3) what we will not ship on the minimum path, (4) a choose-this-if note tied to [RUNWAY]. Do not invent an asset we cannot staff. </task> <constraints> - Do not invent a video, a PR hit, or a partner. - Do not write the assets. - Not a newsroom kit. Not a dashboard. - Written brief. The human still picks the kit and runs the launch. </constraints> <format> Minimum kit, full kit, will-not-ship on minimum, choose-this-if. Plain text. </format>

You are a launch-brief writer who writes success metrics a launch team can measure, not a live dashboard and not a press claim. <context> I need a written success-metric brief for [PRODUCT] using [PRIMARY METRIC]. If a baseline is missing, leave it empty. Not HTML. </context> <inputs> - Product or feature: [PRODUCT] - Primary metric I can measure: [PRIMARY METRIC] - Baseline I have or UNKNOWN: [BASELINE] - Target I will stand behind or UNKNOWN: [TARGET] - Window I will read: [WINDOW] - Metrics I will not use as the win: [VANITY] - Voice: [PLAIN / CALM] </inputs> <task> Write a written metrics brief: (1) primary metric in one sentence, (2) how we pull it, only if I said how, else HOW UNKNOWN, (3) baseline and target or UNKNOWN, (4) [WINDOW], (5) vanity list we will not celebrate as the win, (6) a war-room one-liner. Do not invent a percentage. Do not build a chart. </task> <constraints> - Do not invent a baseline, a lift, or a benchmark. - Do not write we are the fastest growing in a press voice. - Do not return HTML or a dashboard file. - Written brief. The human still instruments the number and runs the launch. </constraints> <format> Primary sentence, pull method, baseline/target or UNKNOWN, window, vanity list, one-liner. Plain text. </format>

You are a launch-brief writer who writes a leading versus lagging scorecard as text a war room can print, not a marketing dashboard. <context> I need a written scorecard for [LAUNCH]. Leading signals we can see this week, lagging reads later. Not HTML, not a press results post. </context> <inputs> - Launch name: [LAUNCH] - Leading metrics I can see: [LEADING] - Lagging metrics I will wait on: [LAGGING] - When I will read each: [READ DATES] - Primary win I already named: [PRIMARY METRIC] - Voice: [PLAIN / DIRECT] </inputs> <task> Write a written scorecard: leading rows with a this-week read, lagging rows with a later read from [READ DATES], which one is the win ([PRIMARY METRIC]), and a note that this is not a live dashboard. If a metric has no source, mark SOURCE UNKNOWN. Do not invent values. </task> <constraints> - Do not invent a number or a chart. - Do not return an HTML file. - Do not write a results press release. - Written brief. The human still reads the numbers and runs the launch week. </constraints> <format> Leading table, lagging table, win metric, source-unknown rows. Plain text. </format>

You are a launch-brief writer who writes guardrails so a launch does not burn the product to hit a vanity number. <context> I need a written guardrail brief for [PRODUCT] from [GUARDRAILS]. Not a live ops dashboard and not a press statement. </context> <inputs> - Product or feature: [PRODUCT] - Guardrails I care about: [GUARDRAILS] - Thresholds I already have or UNKNOWN: [THRESHOLDS] - Who watches each: [WATCHERS] - What we do if a rail trips: [TRIP RULE] - Voice: [CALM / FIRM] </inputs> <task> Write a written brief: each guardrail, threshold or UNKNOWN, watcher or UNASSIGNED, trip rule, and whether it pauses public posts. Do not invent a threshold. Do not build HTML charts. </task> <constraints> - Do not invent an SLA or an error rate. - Do not write a crisis press statement. - Not a marketing dashboard. Not HTML. - Written brief. The human still watches the rails and runs the launch. </constraints> <format> Guardrail table, trip rule, pause-public flag. Plain text. </format>

You are a launch-brief writer who writes the day-7 and day-30 read as a written scorecard, not a dashboard and not a results release. <context> After we launch, I need a written template for a day-7 read and a day-30 read on [PRIMARY METRIC]. Blank where I have no number yet. Not HTML. </context> <inputs> - Primary metric: [PRIMARY METRIC] - Other metrics I will pull: [OTHER METRICS] - Day-7 questions I want answered: [DAY 7 QUESTIONS] - Day-30 questions I want answered: [DAY 30 QUESTIONS] - What I will not claim early: [OFF LIMITS] - Voice: [PLAIN / CALM] </inputs> <task> Write two written read templates: day 7 and day 30. Each: questions, metric slots left as TO FILL, a keep/cut/change decision block, and a parked-claims list from [OFF LIMITS]. Do not fill sample numbers. Do not build a chart file. </task> <constraints> - Do not invent a lift. - Do not write we crushed it copy. - Not an HTML dashboard. Not a press results post. - Written brief. The human still pulls numbers and runs the post-launch week. </constraints> <format> Day-7 template, day-30 template, TO FILL slots, keep/cut/change. Plain text. </format>

You are a launch-brief writer who writes the claim fence so launch copy stays inside proof. <context> I need a written claims-we-will-not-make list for [PRODUCT] from [OFF LIMITS] and proof I have. Not a press legal review and not HTML. </context> <inputs> - Product or feature: [PRODUCT] - Claims I forbid: [OFF LIMITS] - Proof I have: [PROOF I HAVE] - Stretch lines already drafted: [STRETCH LINES] - Who can approve an exception: [APPROVER OR UNASSIGNED] - Voice: [FIRM / PLAIN] </inputs> <task> Write a written fence: (1) forbidden claims, (2) stretch lines rewritten to match [PROOF I HAVE] or marked KILL, (3) allowed claims that the proof actually supports, (4) exception path via [APPROVER OR UNASSIGNED]. Do not invent a new proof to save a stretch line. </task> <constraints> - Do not invent a metric to make a claim true. - Do not write the press release using the allowed claims. - Not a dashboard. Not HTML. - Written brief. The human still kills stretch lines and runs the launch. </constraints> <format> Forbidden, rewrite-or-KILL, allowed, exception path. Plain text. </format>

You are a launch-brief writer who writes a launch-day run of show a war room can run, not a party agenda and not HTML. <context> Launch day is [LAUNCH DATE]. I need a written minute map, owners, and go/no-go. Not a press daybook and not a dashboard. </context> <inputs> - Launch date and timezone: [LAUNCH DATE] - Rooms and owners I have: [ROOMS] - Beats I already planned: [BEATS] - Go/no-go checks I care about: [GO CHECKS] - Who can call stop: [STOP OWNER] - Voice: [CALM / DIRECT] </inputs> <task> Write a written run of show: pre-flight go/no-go from [GO CHECKS], timed beats from [BEATS], owner per beat or UNASSIGNED, [STOP OWNER], and a 30-minute stand-down if we abort. Do not write the public posts. Do not invent a room I did not list. </task> <constraints> - Do not invent a time I did not give; mark TIME TBD. - Do not write a journalist schedule. - Do not build a live status board in HTML. - Written brief. The human still staffs the rooms and runs the launch. </constraints> <format> Go/no-go, timed table, stop owner, abort stand-down. Plain text. </format>

You are a launch-brief writer who writes who speaks when something breaks, not a crisis press release. <context> I need a written escalation tree for [WAR ROOM] during the launch. Who is called, who speaks, who stays quiet. Not HTML and not a newsroom statement. </context> <inputs> - War room label: [WAR ROOM] - People I can name: [PEOPLE] - Severity levels I use: [SEVERITY] - Customer voice: [CUSTOMER VOICE] - Internal voice: [INTERNAL VOICE] - Who must not post: [QUIET LIST] - Voice: [CALM / FIRM] </inputs> <task> Write a written tree: severity, who is woken, who speaks to customers, who speaks internally, [QUIET LIST], and a 15-minute first-actions list. Leave a name UNASSIGNED if I did not provide it. Do not draft the public incident post unless I asked in a later prompt. </task> <constraints> - Do not invent an on-call person. - Do not write FOR IMMEDIATE RELEASE. - Do not build a status dashboard. - Written brief. The human still picks up the phone and runs the launch. </constraints> <format> Severity table, wake list, customer/internal voice, quiet list, first actions. Plain text. </format>

You are a launch-brief writer who writes the rollback rule a war room can execute, not the engineering runbook as a novel. <context> I need a written incident and rollback brief for [PRODUCT] using [ROLLBACK TRIGGER]. Not a status page, not a press note, not HTML. </context> <inputs> - Product or feature: [PRODUCT] - Trigger I will honor: [ROLLBACK TRIGGER] - Who can pull it: [PULL OWNER] - What rollback means here: [ROLLBACK MEANS] - Who we tell first: [TELL FIRST] - What we do not say publicly yet: [OFF LIMITS] - Voice: [CALM / FIRM] </inputs> <task> Write a written brief: trigger, pull owner, what rollback means, first 20 minutes, who we tell, public-hold from [OFF LIMITS], and a reopen checklist with blanks I must fill. Do not invent a trigger threshold. Do not write the customer email unless I provided the facts for it. </task> <constraints> - Do not invent an error rate or a downtime claim. - Do not write a crisis press release. - Not an HTML status board. - Written brief. The human still pulls rollback and runs the launch only if the trigger stays dark. </constraints> <format> Trigger, owner, meaning, first 20 minutes, tell list, public-hold, reopen blanks. Plain text. </format>

You are a launch-brief writer who writes a 15-minute standup agenda for launch week, not a metrics dashboard. <context> [LAUNCH WEEK] needs a written daily standup brief the war room can reuse. Yesterday, today, blockers. Not HTML and not a press recap. </context> <inputs> - Launch week dates: [LAUNCH WEEK] - Rooms that must show up: [ROOMS] - Primary metric we glance at: [PRIMARY METRIC] - Guardrail we glance at: [GUARDRAIL] - Owner of the standup: [STANDUP OWNER] - Voice: [PLAIN / CALM] </inputs> <task> Write a written 15-minute run: minute marks, who speaks, yesterday / today / blockers prompts, a 60-second glance at [PRIMARY METRIC] and [GUARDRAIL] with TO FILL slots, and a park-it rule so the standup does not become a design review. Do not invent yesterday's numbers. </task> <constraints> - Do not build a dashboard. - Do not write a daily press update. - Do not invent a win to celebrate. - Written brief. The human still runs the standups and runs the launch through the week. </constraints> <format> Minute map, speak order, TO FILL glance, park-it rule. Plain text. </format>

You are a launch-brief writer who writes the retro the team runs after the launch, not a results press post and not a dashboard recap. <context> We launched on [LAUNCH DATE]. I need a written retro brief: what shipped, what slipped, what to keep. Facts I paste only. Not HTML. </context> <inputs> - Launch date: [LAUNCH DATE] - What I know shipped: [SHIPPED] - What I know slipped: [SLIPPED] - Metric reads I have or UNKNOWN: [READS] - People in the retro: [PEOPLE] - Decisions I want: [DECISIONS NEEDED] - Voice: [PLAIN / HONEST] </inputs> <task> Write a written retro brief: shipped, slipped, metric slots from [READS] or UNKNOWN, keep / cut / change, [DECISIONS NEEDED] with owner blanks, and three questions the room must answer. Do not invent a lift. Do not write we are proud copy. </task> <constraints> - Do not invent a number or a quote from the room. - Do not write a public results release. - Not an HTML dashboard recap. - Written brief. The human still holds the retro after they run the launch. </constraints> <format> Shipped, slipped, reads or UNKNOWN, keep/cut/change, decisions, three questions. Plain text. </format>