Installable skillMIT

LinkedIn Post

Generate LinkedIn posts about work sessions and technical accomplishments. Use when creating developer-focused content or sharing technical insights.

BY NehaTanti-afkUPDATED Oct 28, 2026
INTENDED FOR✳Claude Code>_Codex◈OpenClaw↗Cursor◎ChatGPT✦GeminiInstall in a listed agent.

instructions.md

---
name: linkedin-post-3389dadcc39f
description: Generate LinkedIn posts about work sessions and technical accomplishments. Use when creating developer-focused content or sharing technical insights.
allowed-tools:
  - Read
  - Grep
  - Glob
  - Bash
user-invocable: true
---

# LinkedIn Post

## Purpose

Generate a LinkedIn post based on the current work session - what was built, changed, or accomplished.

## When to Use

- When user runs `/linkedin` or asks for a LinkedIn post
- After completing a significant feature or task
- To share technical learnings or insights

## Workflow

### Phase 1: Analyze Session Context

**Extract from conversation history:**

1. What task was accomplished?
2. What files were changed/created?
3. What problem was solved?
4. What approach was taken?
5. Any interesting technical decisions?

**Run git commands to understand scope:**

```bash
git diff --stat HEAD~5
git log --oneline -5
```

### Phase 2: Identify the Story

Based on session analysis, identify:

| Element          | Question to Answer                       |
| ---------------- | ---------------------------------------- |
| **The Hook**     | What's the surprising/interesting angle? |
| **The Problem**  | What challenge was being solved?         |
| **The Insight**  | What did you learn or realize?           |
| **The Takeaway** | What can others learn from this?         |

### Phase 3: Draft Post

**Structure:**

```
[Hook - 1-2 lines that stop the scroll]

[Context - What you were working on]

[The interesting part - Technical insight, unexpected learning, or aha moment]

[Takeaway - What others can learn]

[Engagement prompt - Question or call to discussion]
```

**Length:** 150-300 words (LinkedIn sweet spot)

### Phase 4: Self-Review

Apply anti-patterns check:

- [ ] No "Excited to announce" or marketing clichés
- [ ] No buzzwords without substance
- [ ] Has a clear hook in the first line
- [ ] Creates insight, not just description
- [ ] Invites conversation
- [ ] Appropriate technical depth for LinkedIn audience

### Phase 5: Present Options

Offer the user:

1. **Draft post** - Ready to copy
2. **Alternate angles** - 2-3 different hooks/approaches
3. **Revision requests** - Ask what to adjust

## Output Format

```markdown
## 📝 LinkedIn Post Draft

{post content}

---

### Alternative Angles

**Option A:** {different hook focusing on X}
**Option B:** {different hook focusing on Y}

---

**Want me to:** Adjust tone? | Add more detail? | Make it shorter? | Try a different angle?
```

## Dynamic Content Sources

The skill derives ALL content from:

| Source                | What to Extract                    |
| --------------------- | ---------------------------------- |
| Conversation history  | Task accomplished, problems solved |
| Git diff/log          | Files changed, commit messages     |
| File contents read    | Technical details, code patterns   |
| User's explicit input | Any additional context provided    |

**Never hardcode** project names, tech stack, or domain assumptions - discover them from the session.

## Anti-Patterns to Avoid

- "Excited to announce" and similar marketing clichés
- Buzzwords without substance (synergy, leverage, cutting-edge)
- Posts that only describe without creating insight
- Missing hook in the first line
- No clear call to engagement
- Too technical for the platform audience
- Too vague to be interesting
COMPATIBILITY & REFERENCES

Platform guidance

Imported from the listed AgentSkill.sh category. Complete supplied package files and notices are retained. Required external tools and credentials are not installed. No bundled code was executed during import. Individual agent task behavior has not been tested.