Thought Leadership

Doctrine inherited (Step 7 — 0626 rollout)

Output complies with:

• output-tenets.md — the seven tenets (Storytelling/SQCA is Tenet 7)
• output-simplicity.md — length caps, three-layer source placement, robot-tells ban
• ai-speak-anti-patterns.md — the 12 patterns thought-leadership hooks can't carry
• Step 6 calibration: see [[feedback_execution_doctrine_refinements_step6]]

Refinements applied to this skill:

Code	Refinement	How it lands in thought-leadership
R1	Source placement (three layers)	Published essay is end-customer-facing. No [VERIFIED:...] tags inline. Citations convert to inline links inside the prose at publish time (e.g., "Tan's post on Series A spread"). Working draft carries internal [VERIFIED:...] for the phase-1 → phase-2 review gate, stripped before phase 3 polish.
R3	Product-update tone	When the essay references the author's product, frame as "we ship X" not "we are thrilled to announce." The contrarian angle does the work — adjectives don't.
R5	Blog as voice anchor	Long-form essay IS the voice anchor. Downstream LinkedIn / newsletter / podcast assets paraphrase forward from the essay's opening line — that's where R5 originates upstream.
R9	Action-oriented section names	Section headers state the claim (not "Background / Conclusion"). "Why the spread game is broken" beats "The Series A problem." Claim-led, not category-led.

When to run

Long-form articles, essays, whitepapers, or guest posts where the user has a contrarian or non-obvious angle they want to publish. Three human-approval gates (Phase 1 pyramid, Phase 2 prose, Phase 3 polish). Skip for LinkedIn posts (linkedin-content), newsletter copy, or website copy. Triggers, anti-triggers, and pre-flight checklist → the premium reference.

Inputs

Input	Status	Source
Topic / core insight	Required	User — must be specific and contrarian
Target audience	Required	ICP research or user-specified
Core argument	Required	User or collaboratively developed
ICP research	Recommended	clients/[client]/icp-behavioural.md
Product messaging	Recommended	clients/[client]/product-messaging.md
Content strategy	Recommended	Pillars + roadmap
Competitor positions	Recommended	Identifies white space
Source material, word count, publication context	Optional	LinkedIn 800-1500 / blog 1500-2500 / guest 2000-3500

Client CLAUDE.md is auto-loaded — apply its voice/messaging anchors. Full input matrix and pre-flight checklist in the premium reference.

Steps

1. Validate inputs — topic specific, audience known, contrarian angle present, user has time for 3 review cycles.
2. Phase 1.1 — Clarify core insight. Ask: "What's the one thing you want the reader to believe after reading this that they didn't believe before?"
3. Phase 1.2 — Identify audience starting point. Current belief, gap to governing thought, anticipated objections.
4. Phase 1.3 — Build the pyramid. Governing Thought → 2-4 Key Line Arguments → Supporting Evidence per key line. Use the template in the premium reference. Pyramid theory + MECE deep-dive in the premium reference.
5. Phase 1.4 — Test structure. MECE check (no overlap, fully supports governing thought) + "so what?" test on every level.
6. Phase 1 deliverable + STOP. Wrap in the Phase 1 checkpoint format. Wait for explicit "Approved — proceed to Phase 2" before continuing.
7. Phase 2.1 — Map pyramid to prose sections. Governing thought → opening, key lines → sections, objections → inline or dedicated, "so what?" → closing.
8. Phase 2.2 — Write opening. Hook → Context → Thesis → optional Roadmap. Phase 2.3 — Expand each section: Headline → Lead → Evidence → Transition. Phase 2.4 — Write closing: Restate → Synthesize → Implications → CTA.
9. Phase 2 deliverable + STOP. Three title options + full prose outline. Wrap in Phase 2 checkpoint format. Wait for "Approved — proceed to Phase 3".
10. Phase 3.1 — Apply TOV. Operator authority, prescriptive confidence, framework-driven clarity, candid partnership, specificity over superlatives, conversational warmth. Em dashes with spaces, sentence case headlines, numerals, banned-word check. Quick checklist → the premium reference.
11. Phase 3.2 — ICP guardrails. Would target ICP care? Right sophistication? Would they share this with their team?
12. Phase 3.3 — Anti-hallucination guardrails. No invented client names/metrics/quotes. Mark missing data as [Data point needed]. Verify proof points or flag as [Example — verify before publishing].
13. Phase 3.4 — 100 Posts Test. "If I published 100 posts like this, would the aggregate feel authentic to my voice and brand?" Plus final read-aloud check.
14. Phase 3 deliverable + final STOP. Wrap as THOUGHT LEADERSHIP FINAL — PHASE 3 with quality report and items-for-review. Wait for "Approved — ready to publish".
15. Push to chains. After approval, offer LinkedIn teaser, standalone quote pulls, email intro, or newsletter summary. Suggest /linkedin-content, /email-nurture, or update /content-strategy plan.

Format guidelines per platform → the premium reference. Phase output wrappers, checkpoint dialogues, MECE details, Exa research substrate, and full auto-update protocol → the premium reference.

What good looks like

Examples — examples/worked-example-gtm-engineer.md (full 3-phase walkthrough), examples/top-performing-examples.md (validated patterns).

Evaluations — Governing thought is contrarian and specific (not commodity wisdom). Key lines are MECE. Every claim has verified source or [Example — verify] flag. No banned words: solutions, leverage, synergy, scalable, best-in-class, game-changer, innovative, cutting-edge, very, really, basically, just. Three checkpoints honoured (no skipping ahead). Passes 100 Posts Test. Cites [VERIFIED: exa_search, {url}, accessed {YYYY-MM-DD}] for any external claim per .claude/rules/ontology.md and .claude/rules/exa-protocol.md.

Final ship gate

Run /premortem --output before ship. See /premortem skill for the 5 execution domains (will-it-resonate / will-it-convert / will-it-stay-on-brand / will-stakeholder-push-back / will-it-degrade-over-time) and output template.

Then run /voice-reviewer — the content ship gate: voice + brand quality (pm-loop.md).

Trivial-case escape: ## Premortem\nNo failure modes — trivial change satisfies the contract for genuinely trivial outputs.

---

Persuasion & stickiness pass

Output complies with persuasion-and-stickiness.md — Cialdini's 7 persuasion levers + Heath's SUCCESs. Deploy the 1-2 Cialdini levers that fit the reader's barrier (never all seven; every lever must be TRUE), run the SUCCESs diagnostic (Simple / Unexpected / Concrete / Credible / Emotional / Stories) over the near-final draft, then the rule's pre-ship gate.