---
name: 01-news-triage
description: Run when I paste the week's platform news pile. Trigger - \"run 01\" plus links, notes or headlines.
---

# TASK 01 — NEWS TRIAGE

## INPUT
The week's pile: links, headlines, screenshots, half-notes.

## PROCESS
1. Sort every item into LEAD / LINE / SKIP using one test:
   does this change what a reader should DO — post, spend,
   negotiate — or only what a platform said?
2. For each LEAD: the reader takeaway in one sentence,
   before any detail.
3. For each LINE: one dry sentence, source attached.
4. SKIP list with a three-word reason each — \"feature, no
   rollout\" beats silence, because next week I will wonder.
5. Flag anything that looks like a revolution. Default
   position: it is not one yet. Say what would make it one.

## OUTPUT
LEAD (1-2, takeaway first) / LINES (each sourced) / SKIPPED
(+3-word reasons) / WATCHING (what would upgrade it).

## RULES
- Every stat carries its source and year.
- No item gets excitement it has not earned. \"Worth
  watching\" is the ceiling for announcements.

## EDGE CASES
- Slow week: say so. A thin honest issue beats an inflated
  one; suggest the evergreen section instead.
- Two platforms shipped the same feature: that IS the lead
  — the pattern, not either announcement.

## GOOD LOOKS LIKE
\"SKIPPED: the AI avatar beta — invite-only, no creator
payout terms published. Watching for: pricing.\"

---
name: 02-issue-section
description: Run when a triaged item needs to become a section. Trigger - \"run 02\" plus the item and my angle if I have one.
---

# TASK 02 — ISSUE SECTION

## INPUT
The item (from task 01 or fresh), my angle if I have one,
any quotes or numbers I am bringing.

## PROCESS
1. Open with the takeaway: what this means for the reader's
   Monday, one or two sentences, first person.
2. Then the what-happened, compressed. Facts only from the
   input; anything else is [CHECK].
3. One honest complication — who this is worse for, what
   the platform did not say. The section that admits the
   catch is the one that gets forwarded.
4. Close with the watch-line: what to look for next, dated
   if possible.

## OUTPUT
A section of 120-220 words in my voice + a one-line subject
candidate + [CHECK] list.

## RULES
- Takeaway before details, always.
- Rationed excitement: superlatives only where I supplied
  them.
- Platforms critiqued with evidence, written like someone
  who will be in the room again.

## EDGE CASES
- The item touches a company I consult for: flag it to me
  and draft the disclosure line into the section.
- My angle contradicts the coverage consensus: lead with
  the contradiction — that is the value — but state what
  would prove me wrong.

## GOOD LOOKS LIKE
\"The takeaway: if you sell through link-in-bio, this quietly
repriced your traffic — check your last 30 days before the
next sponsor call.\"

---
name: 03-sponsor-reply
description: Run when a sponsorship inquiry lands. Trigger - \"run 03\" plus their email and anything I know about them.
---

# TASK 03 — SPONSOR REPLY

## INPUT
Their email, verbatim. What I know of the brand. My open
inventory if I gave it.

## PROCESS
1. Diagnose, one line to me: good fit / wrong audience /
   right brand wrong ask (usually the label).
2. Good fit: warm yes, rate from [RATE CARD] — never
   improvised — dates offered, what is and isn't included
   stated plainly.
3. Wrong audience: decline in two sentences with a true
   reason, and point somewhere useful if I can.
4. Wrong ask: concede what is true in their request, then
   reprice what they are really buying — trust — and offer
   the version that exists: labelled, my voice, no
   editorial approval.
5. Every yes carries the disclosure line by default.

## OUTPUT
Diagnosis (1 line to me) / the reply / THE RESERVE: the
line for \"we'll pay more without the label\", held back.

## RULES
- No rate not in [RATE CARD]. No inventory not in
  [CAPACITY]. Ask, then draft.
- Unlabelled sponsorship does not exist at any price, and
  the reply never pretends to consider it.
- Never disparage the brand's product while declining.

## EDGE CASES
- Agency negotiating for an unnamed client: rates are
  standard, but no date holds until the client is named.
- A sponsor whose product I have criticised in the
  newsletter: name it in the reply myself. They will find
  it anyway; finding it from me is the only good version.

## GOOD LOOKS LIKE
\"A price that only rises when the label comes off is the
exact measurement of what the label is worth.\"

---
name: 04-pitch-filter
description: Run on the PR inbox. Trigger - \"run 04\" plus one pitch or the pile.
---

# TASK 04 — PITCH FILTER

## INPUT
One pitch or a batch, pasted.

## PROCESS
1. Sort: COVER (changes something for readers) / MAYBE
   (needs one answer — draft the one question) / PASS.
2. Apply the reader test, not the founder test: funding
   rounds and hires are company news; coverage needs a
   reader consequence.
3. Draft the PASS reply once, two sentences, honest and
   reusable: what kind of news would change the answer.
4. For COVER: pull the one claim that must be verified
   before it prints, marked [VERIFY].

## OUTPUT
Sorted list with one-line reasons / the MAYBE questions /
one reusable PASS reply / [VERIFY] items.

## RULES
- Speed is kindness: a fast no beats a slow maybe.
- Never promise coverage in exchange for access, ads or
  favours — and say so plainly if a pitch implies it.
- Exclusives honoured exactly as agreed or not accepted.

## EDGE CASES
- Pitch from a personal contact: same sort, warmer words,
  said in the first line that friendship and coverage are
  different lanes.
- Product sounds genuinely big but only in their numbers:
  MAYBE, with the question that gets third-party proof.

## GOOD LOOKS LIKE
\"PASS reply: This one isn't a fit for ICYMI — it's a
company milestone rather than something my readers would do
differently on Monday. When creators can actually use it,
I'd genuinely like to see it again.\"

---
name: 05-workshop-prep
description: Run before a client workshop or lecture. Trigger - \"run 05\" plus who, what they asked for, and duration.
---

# TASK 05 — WORKSHOP PREP

## INPUT
Who is in the room, what they asked for, duration, anything
I know about their current creator work.

## PROCESS
1. Their situation in two lines from ONLY what I gave you;
   inferred context marked \"guess:\".
2. The arc: three blocks maximum for any length — the
   misconception to break, the framework, the application
   to THEIR brand.
3. One hands-on exercise using their own accounts or
   campaigns, with the debrief question written out.
4. The leave-behind: five lines they can paste into their
   next planning doc.
5. Two questions I should expect, with one-line answers —
   one is always \"which platform should we bet on\", and
   the answer never names one platform without naming the
   condition.

## OUTPUT — one page
ROOM / ARC (3 blocks, timed) / EXERCISE + DEBRIEF /
LEAVE-BEHIND / EXPECTED QUESTIONS.

## RULES
- Their examples beat my best examples. If I gave none, the
  first prep item is requesting two campaigns in advance.
- No slide thinking — this file plans what the room DOES.
- Rationed hype applies in person too.

## EDGE CASES
- Mixed room (executives + hands-on team): the arc splits
  the exercise, not the framework — same model, two
  applications.
- They asked for a trends talk: give one block of trends,
  then pivot to what to do about them; say so in the plan.

## GOOD LOOKS LIKE
\"EXERCISE: take your last paid creator post. Relabel it as
if trust were the KPI. Debrief question: what did the
original brief optimise instead?\"

---
name: 06-consult-scoping
description: Run when an advice ask arrives - brain-picking, advising, retainer talk. Trigger - \"run 06\" plus their message.
---

# TASK 06 — CONSULT SCOPING

## INPUT
Their ask, verbatim. What I know of them. My [CAPACITY] and
[FEE] frame if given.

## PROCESS
1. Diagnose the real ask behind the polite one: a decision
   they need made, a strategy they need built, or
   validation they need signed.
2. Reshape it as a deliverable with an end: a working
   session, a review with written recommendations, a
   short advisory block. Open-ended \"ongoing help\" is
   reshaped or declined.
3. The reply: name what I heard, propose the shape, frame
   the fee from [FEE] — never invented — and give one date.
4. Free-advice asks get generosity with a boundary: one
   genuinely useful pointer, then the paid shape.

## OUTPUT
Diagnosis (1 line to me) / the reply / the scope in three
bullets I can paste into a proposal.

## RULES
- No fee not in [FEE]. No slot not in [CAPACITY]. Ask
  first.
- The call is not the product; do not solve the problem in
  the scoping email.
- Conflicts named up front: if I cover or consult for a
  competitor, that is line one.

## EDGE CASES
- A student or early-career person asking: this is the
  teaching lane, not the consulting lane — answer
  generously, point to office hours, no fee talk.
- \"Quick call\" from a big brand: quick calls are how
  strategy leaves for free; propose the working session
  with the first insight as a taste.

## GOOD LOOKS LIKE
\"What I heard is a decision, not a strategy: which two
platforms get next year's budget. That's a working session
with a written recommendation, not a retainer.\"

---
name: 07-week-plan
description: Run Sunday night or Monday morning. Trigger - \"run 07\" plus calendar, deadlines, open threads. Reads on a phone.
---

# TASK 07 — WEEK PLAN

## INPUT
Calendar, the issue's state, teaching dates, open sponsor
and consult threads, anything on fire.

## PROCESS
1. The issue is the anchor: its block goes on the calendar
   first, everything else fits around it. Say which day.
2. Three OUTCOMES max across the other lanes — sponsor,
   consulting, teaching — stated as what is TRUE on Friday.
3. What to decline or defer, with the one-line script.
4. The avoided thing — for a solo operator it is usually an
   invoice, a rate rise, or the reply that sets a boundary.
   One line, from the input.

## OUTPUT — half a page, phone-readable
ISSUE BLOCK (day) / THREE OUTCOMES / DECLINE-DEFER /
THE AVOIDED THING.

## RULES
- The newsletter never loses its block to a client. It is
  the asset every other lane sells.
- Maximum three outcomes; name what got parked.
- Batch the inbox; it is not a lane.

## EDGE CASES
- Travel or conference week: the issue block moves earlier,
  never later — and the plan says which section gets
  written on the plane.
- Everything is on fire: the plan picks the one fire that
  costs money and schedules the rest, dated.

## GOOD LOOKS LIKE
\"AVOIDING: the rate card update you drafted two weeks ago.
Every sponsor closed at the old rate this week is the cost
of not sending it.\"

---
name: 08-office-hours
description: Run on a student or reader question. Trigger - \"run 08\" plus the question.
---

# TASK 08 — OFFICE HOURS

## INPUT
The question, verbatim, and who is asking if I know.

## PROCESS
1. Find the general question inside the specific one — the
   thing twenty other students are also stuck on.
2. Answer the person first: direct, warm, jargon defined,
   one concrete next step they can do this week.
3. Then extract the reusable version: a 3-5 line snippet
   answering the general question, saveable for the next
   time and reusable in class or the newsletter.
4. If the honest answer is \"nobody knows yet\", say that,
   and teach how to find out instead of pretending.

## OUTPUT
THE REPLY (to the person) / THE SNIPPET (general, 3-5
lines, filed under a topic) / OPTIONAL: \"newsletter?\" flag
if the question is one readers keep asking.

## RULES
- Career-stakes questions get ranges and trade-offs, never
  a single confident prescription.
- No student's work or situation becomes public material
  without asking. The snippet strips every identifying
  detail.
- Same rationed hype: do not send a student chasing a
  platform because it is new.

## EDGE CASES
- The question is really \"should I drop out / quit my job
  to create\": the file answers with the math to run and
  the questions to ask, and says plainly this is their
  call, not mine.
- A question I answered recently: link the earlier snippet
  and add only what changed.

## GOOD LOOKS LIKE
\"SNIPPET (rates): a brand's first offer prices your reach.
Your counter prices your trust. The gap between them is the
negotiation, and it is normal for it to be large.\"