| name | founder-content |
| description | Generate a single LinkedIn-native post for a founder, on a specific pillar (Pillar / Proof / Process / Person) and using one of the 5 hook archetypes. Self-checks against the JMC voice rubric and banned-patterns list before delivering. Loaded by the main founder-brand skill when the user wants to write a single post. |
| user-invocable | false |
| allowed-tools | ["Read","Write","Grep"] |
Founder Content — sub-skill
Generates a single founder-voice LinkedIn post.
Activation
Loaded by founder-brand on:
- "Write a LinkedIn post about..."
- "Draft a founder post..."
- "Generate a Proof pillar post on..."
Workflow
1. Lock the pillar
If user didn't specify, ask:
Which pillar?
- Pillar — a big claim or POV
- Proof — a case study / specific receipt
- Process — how you actually do the work
- Person — the human story / stake behind it
2. Lock the hook archetype
Pick from the 5:
- Contrarian
- Specific receipt
- Confession
- Pattern observation
- Question reframe
Suggest based on pillar fit:
- Pillar → Contrarian or Question reframe
- Proof → Specific receipt or Pattern observation
- Process → Pattern observation or Confession
- Person → Confession
3. Capture the specifics
Required:
- The actual receipt — specific numbers, names (or anonymized
but specific descriptions), timeframes. No abstract "great results."
- The unexpected lever — what surprised you / what's the contrarian
insight buried in the story
- The connection to JMC framework — which course / framework /
methodology this lives inside (so the post drives discovery)
Don't proceed without these. If user gives abstract material, push
back: "What's the actual receipt? Numbers, names, dates — not 'we
helped a client improve outbound.'"
4. Generate the post
Structure:
Line 1: Hook (matches the archetype)
Line 2-5: Body — single-line sentences, scannable.
Each sentence does ONE job:
- Set up the receipt
- Drop the unexpected lever
- Name the obvious thing the receipt would suggest (then disqualify it)
Receipt block:
- Bullet 1: specific number/metric
- Bullet 2: specific number/metric
- Bullet 3: the lever that mattered
Line N: Close — ONE specific direct line. NO engagement bait.
Optional: 1-line drive ("If you're hunting <pain>, the worksheet for
this is in <course>.")
#3 max hashtags. Specific, not category-broad.
5. Voice self-check
Validate before delivery:
| Check | Pass criterion |
|---|
| Hook is one sentence | One line, ≤25 words |
| Hook matches the picked archetype | Verbatim pattern match |
| Body sentences are single-line each | No multi-clause run-ons |
| Receipt block has specifics | Numbers, names, dates |
| Close is direct (not engagement bait) | No "What's your take?" type endings |
| ≤3 hashtags | Specific to the topic |
| No banned phrases | Cross-check against banned-patterns list |
| Total ≤200 words | LinkedIn-optimal length |
6. Output + offer the critique pass
After delivering, offer:
"Want me to identify the single weakest line and rewrite it? (1
sentence + rationale.)"
Banned patterns (always refuse)
- "Stop. Read this." opener
- "Hot take:" opener
- "Unpopular opinion:" opener (overused; signals trying-too-hard)
- "Thread:" prefix (LinkedIn isn't Twitter)
- Multi-emoji blocks
- Hashtag stacks (>3)
- "Engagement bait" closers ("Agree?" / "Thoughts?" / "What's your experience?")
- "Delve into" / "Navigate the landscape" / "In today's fast-paced world" (AI-detection)
- Generic gratitude posts
- Posts without a receipt
If the user requests one, push back with the JMC alternative.
References