borghei/Claude-Skills / tools/linkedin/linkedin-humanizer
linkedin-humanizer
VerifiedAudits and rewrites LinkedIn drafts to remove machine-sounding patterns: tiered tell catalogue, emoji-pattern scoring, rule explanations, and a voice fingerprint built from the author's own posts. Use when a draft reads generated, before publishing, or when an edit flattened someone's voice.
| File | Kind | Size |
|---|---|---|
| SKILL.md | skill md | 17.9 KB |
| assets/audit_report_template.md | markdown | 2.3 KB |
| assets/sample_draft.txt | markdown | 1.0 KB |
| assets/sample_draft_edited.txt | markdown | 1.1 KB |
| assets/sample_past_posts.txt | markdown | 2.7 KB |
| assets/sample_voice.json | config | 899 B |
| assets/voice_profile_template.md | markdown | 3.0 KB |
| references/emoji-patterns.md | markdown | 7.4 KB |
| references/rewrite-playbook.md | markdown | 12.0 KB |
| references/rule-catalogue.md | markdown | 19.4 KB |
| references/voice-fingerprint-guide.md | markdown | 9.7 KB |
| scripts/emoji_audit.py | script | 8.6 KB |
| scripts/tell_audit.py | script | 10.8 KB |
| scripts/tell_rules.py | script | 17.5 KB |
| scripts/voice_fingerprint.py | script | 11.1 KB |
assets/sample_past_posts.txt
View on GitHubWe closed 1,140 support tickets in August with five people. In August last year it took nine people to close 900.
I'd love to say that was a tooling win — it wasn't, mostly. Jonas rewrote our 14 most-used macros so they answered the question in the first line instead of the third, and that's most of it.
The part I didn't expect: reopens dropped too. People don't write back when the first reply actually answers them.
If your queue is growing, read your ten most-used macros out loud before you buy anything.
---
I got a promotion wrong last spring and it took me four months to notice.
Priyanka was the best agent on the team — fastest, kindest, the one customers asked for by name. So I made her a team lead. By July she'd stopped taking tickets, hated her calendar, and two of her reports had asked to move.
Nothing was wrong with her. I'd taken the thing she was great at and replaced it with meetings.
We split the role in September. She's a senior agent again with a pay band that matches the lead track, and she runs the Thursday review, which she likes. Nobody's asked to move since.
I'd assumed the only way up was management because that's how it worked everywhere I'd been. It doesn't have to.
---
A customer told me last week that our status page was "technically honest and practically useless."
She was right. It said "degraded performance" for six hours on the 3rd while label printing was fully down for anyone on the old API.
We've changed two things. Incidents now name the feature that's broken in plain words, and whoever's on call writes the update, not comms. It's rougher. It's also true — and it's posted about twenty minutes sooner.
I still don't know how to show partial outages well. If you've seen a status page that does, send it my way?
---
Small thing that's saved me hours: I stopped writing "let me know if you have questions" at the end of handover docs.
Now the last line is the one question I think the reader will have — and the answer.
Tried it on the warehouse onboarding guide in June. Follow-up messages on that doc went from eleven the month before to three.
It works because I have to actually guess what's unclear, and usually that means I go back and fix the unclear bit.
---
We turned off chat for nine days in October. On purpose.
The team was drowning — median first reply had crept to 47 minutes, and half the chats were people asking where a parcel was, which the tracking page already answers.
So chat went away and a bigger "Where's my parcel?" box went on the help page. Email volume went up by about a fifth. Total contacts went down by a third.
We've turned chat back on, but only for accounts with an open incident. I thought customers would be angry. Four complained. One of them was my mum, who doesn't use the product.
Related skills
| Skill | Stars | Updated | Works with | Status | Install |
|---|---|---|---|---|---|
| rea-analysis-optimizationmorluto/reaOptimize REA providers and shared analysis workflows for agent context, latency, CPU, and memory while preserving correctness and Evidence. Use for analysis bottlenecks and cache or query design; CI/test-suite optimization belongs to rea-ci-tests-audit. | Stars: 472k | Updated | Claude CodeCodex | Verified | Install |
| rea-ci-tests-auditmorluto/reaAudit REA CI and test suites for correctness, fixture drift, false passes, and measured performance; use for suite or workflow investigations and efficiency changes, rather than ordinary feature testing or a single test failure. | Stars: 472k | Updated | Claude CodeCodex | Verified | Install |
| rea-tool-designmorluto/reaAudit, investigate, design, or change REA investigation tools, CLI/MCP contracts, provider capabilities, and Evidence semantics. | Stars: 472k | Updated | Claude CodeCodex | Verified | Install |
| ponytail-auditDietrichGebert/ponytailQuality audit of the whole repo: bugs, security, real load, missing tests, speed, and what to delete. Most important first. | Stars: 194k | Updated | Claude CodeCodex | Verified | Install |