founder-report

name	founder-report
description	Founder-report packaging stage for the Diffmode growth-tactics pipeline — the post-gate step that turns the APPROVED synthesis.md (a reviewer-facing working paper) into growth-tactics.md, the founder-facing "Your Growth Tactics" report. Strict re-packaging per the central writing-style.md — plain tactic cards (what it is, why they won't see it coming, how to start this week, time to first signal, what you need, kill it if) with ZERO pipeline plumbing — no vector IDs, no Pass labels, no pools, no numeric scores, no self-grading. Adds no new claims, numbers, tactics, or estimates — every sentence traceable to synthesis.md. Use after the synthesis build reviewer gate and before the HTML report render. Best-effort by design — if it fails, the report falls back to synthesis.md.
metadata	{"version":"1.0.0"}

Founder Report — package synthesis.md into "Your Growth Tactics"

synthesis.md is written to pass a reviewer: it carries disposition tables, traceability lines, pool audits, and self-grades, because downstream checks parse them. That makes it a working paper, not a deliverable. This stage derives the founder view: it reads the APPROVED synthesis.md and re-packages the tactics into growth-tactics.md — the page the founder actually opens first. synthesis.md itself is never modified (its structure is a load-bearing contract with the reviewer and the orchestrator's checks).

The one rule — re-package, never re-author

Zero new content. Every sentence in the output must be traceable to something synthesis.md already says. Specifically:

• No new claims, numbers, estimates, channels, or tactics. If synthesis.md doesn't say it, the report doesn't either.
• Tactic count and names are preserved exactly — same tactics, same order, same names — minus bracket tags: strip the — [Pass 1 White Space] / — [Pass 2 Synergy/Founder-Fit] origin labels from titles, and strip [HIGH COMPLEXITY] / [LOW CONFIDENCE] flags from names. Where such a flag exists, say it in plain words inside the card instead (e.g. "this is the heaviest lift in the set" under What you need).
• Allowed transformations: reorder, merge fields, strip labels and citations, split long sentences, and simplify wording to meet writing-style.md. Nothing else.

Inputs & Output

• WS/03-think-tanks/demand-generation/synthesis.md (PRIMARY — the APPROVED final synthesis; read-only)
• ${CLAUDE_PLUGIN_ROOT}/reference/writing-style.md (the voice contract — grade 6–8, ≤~20 words per sentence, banned-jargon table, the coffee test)
• OUTPUT: WS/03-think-tanks/demand-generation/growth-tactics.md (write no other files)

Document skeleton (exactly these sections, in this order)

1. # Your Growth Tactics

A 2-3 sentence intro: what these tactics are, why they fit THIS founder (pull the budget / team / stage facts from the Synthesis Overview's Constraint Summary), and how to use the document (start with the quick wins, read each card before committing).

2. ## Start this week

The tactics whose first signal arrives within ~4 weeks — take them from the Tactical Clusters quick-wins group. One line each: the tactic name + what you'd do first. No card detail here; the cards follow.

3. ## The tactics

One card per tactic, in synthesis order. Heading: ### N. <Name> (sequential number, plain name, no bracket labels). Each card has exactly these 7 fields, in this order:

Field	Built from (synthesis.md)	Rules
What this is	In plain English (+ Emergent Mechanism where it clarifies)	1-3 short sentences; what you actually do
Why they won't see it coming	Structural Advantage (+ the competitive-adoption note from Primary Channel(s))	lead with why competitors won't think of this themselves; strip taxonomy labels ("Type: …") — keep the plain structural reason and how long the edge lasts. KEEP competitor-adoption lines like "nobody in your market does this yet"
How to start this week	How to Execute steps, kept verbatim where they're already plain	fold in useful Adapted to Your Constraints context (e.g. "scaled down to fit your $500/month")
Time to first signal	Expected Timeline & Success Metrics	the early signal + when to expect it; strip any (per …) vector citations
What you need	Required Resources (+ Skills Required where it adds something)	time / budget / tools, in one or two lines
Kill it if…	the decision point in Expected Timeline & Success Metrics	the concrete number/date that says stop or double down
Execute this	derived from the tactic's first steps	2-3 specific tools, skills, or templates that help execute the first steps — see rules below

"Execute this" rules. This field is the ONE exception to the "zero new content" rule — it names execution aids, not strategic claims. Scope it tightly:

• Name 2-3 concrete tools, agent skills, or templates that help the founder execute the tactic's first steps. Match them to what the tactic actually involves:
  • Outreach / cold email → the cold-email skill (coreyhaines31/marketingskills) or the agent's built-in writing
  • Copy / messaging → the copywriting skill or the agent's built-in writing
  • Landing page / demo → v0.dev, the free-tools skill, or the agent's code generation
  • Community / social → the community-marketing or social skill
  • Content / SEO → the content-strategy or ai-seo skill
  • Design / visual assets → Canva, Figma, or the agent's image generation
  • Tracking / measurement → a simple spreadsheet with columns for the kill-signal metric
• Format as a short bulleted list: → <what>: <tool/skill name>.
• Do NOT invent tactics, channels, or numbers here — only name execution tools.
• If the tactic is purely manual (e.g. "walk into 10 coffee shops"), say → No tools needed — this is a shoes-on-the-ground play.

4. ## What to avoid

From the Anti-Portfolio. Per item: why founders try it, why it fails for THIS founder, and what to do instead (point at the relevant card by number). No severity labels, no "anti-pattern" vocabulary — just plain warnings.

5. ## How to sequence these

The timeline clusters in plain words (this week / this month / the long game), then the Top 5 as "if you only run five, run these" — each with a one-line plain reason taken from what synthesis.md already says about it (fit, speed, the edge). No numeric scores anywhere — the founder gets reasons, not grades.

6. ## Where these came from

One short paragraph: these tactics were built by combining growth mechanisms mined from real, public case studies during this run, and checked against this founder's budget, team, and stage; the reason they're defensible is that copying them would force a bigger competitor to work in a way their structure doesn't allow. Then point the curious reader at the two working papers in the same report: How These Were Built (the mechanism combinations) and Tactic Engineering Notes (the full engineering write-up, scores and traceability included). This must be the LAST section.

Banned in the output (deterministic — the orchestrator greps for these)

The orchestrator rejects the output if any of these appear. Self-grep before returning:

Banned	Why
vector IDs — regex (struct|lever|resource|psych|pos|conv)-[0-9]	plumbing; lives in synthesis.md only
Pass 1 / Pass 2	pipeline phase labels
must_include	constraints plumbing
Pool A / Pool B	constraints plumbing
white space / white-space (any case)	pipeline vocabulary
growth-factors.json	internal file name
PASS / FAIL self-grading	the founder report doesn't grade itself
/10 or /50 scores	no numeric scores in the founder view
"anti-pattern" (any case)	jargon; say what to avoid in plain words

Plus everything writing-style.md already bans (leverage, optimize, arbitrage, …).

Self-check (before returning)

• Card count under ## The tactics == the number of ### Tactic #N blocks in synthesis.md, same order, same names (minus bracket tags).
• Every card has all 7 fields, in the listed order (including "Execute this").
• Banned-pattern grep over the output is empty.
• The document's last section is ## Where these came from.
• Voice: grade 6–8, ≤~20 words per sentence, passes the coffee test, no banned jargon — per writing-style.md.
• Spot-check 2-3 cards: every sentence traces to the matching tactic in synthesis.md — no invented numbers, channels, or steps.

This stage is best-effort by design: if it can't produce a clean output, the pipeline ships synthesis.md as the main page (old behavior). Never degrade synthesis.md to make this stage pass.