| name | cold-email-subject-lines |
| description | Generate or critique cold-email subject lines across 4 framework families (pain, curiosity, social proof, direct). Each suggestion is ≤7 words, no spam triggers, no clickbait, includes reply-rate notes from the JMC Subject Line Swipe File (25 high-performers). Loaded by the main cold-email skill when the user asks for subject lines, openers, or to critique a draft subject. |
| user-invocable | false |
| allowed-tools | ["Read","Grep"] |
Cold Email Subject Lines — 4-Framework Pattern Library
Loaded by cold-email when the user wants subject-line suggestions or
critique.
Activation triggers
- "Subject lines for..."
- "Subject line ideas"
- "Critique this subject"
- "Better subject than..."
- "What's a good subject for..."
The 4 framework families
| Framework | When to use | Example |
|---|
| Pain | Audience is problem-aware | "Pipeline gap after the Q3 hire freeze?" |
| Curiosity | Audience needs a hook before they'll open | "What tried after their Series B" |
| Social proof | Audience trusts other companies more than vendors | "How hit 40% reply rate" |
| Direct | Audience is short on time, you have a clear ask | "15 min Thursday re: demand gen?" |
The skill's job: given a draft cold email (or a target audience + EVP),
suggest 3-5 subject lines spanning at least 2 frameworks.
Constraints (every subject)
- ≤7 words (Gmail truncates around 30 chars mobile)
- No ALL CAPS
- No emojis (deliverability)
- No
Re: / Fwd: fakes (banned by bulk-sender rules)
- No clickbait ("You won't believe...", "This one trick...")
- No spam triggers — load
references/spam-trigger-words.md for the list
- Personalization (first name, company) where relevant — but not faked
- Lowercase first letter often outperforms capitalized in B2B (test)
Workflow
1. Gather context
| Input | Why needed |
|---|
audience (PSP description) | Determines which framework family fits |
evp | Anchors social-proof and pain framings |
intent (open / reply / book) | Direct frameworks work better for booking |
relationship (cold / re-engagement / referred) | Different opening patterns |
2. Generate 5 suggestions
Default: 1-2 per framework family (3-5 total). Order by predicted
fit, not framework alphabetical.
| # | Framework | Subject | Notes |
|---|---|---|---|
| 1 | Pain | "Pipeline gap after the freeze?" | Anchors PSP directly; for problem-aware audience |
| 2 | Social proof | "How <peer> hit 40% reply rate" | Best if you have the case study |
| 3 | Curiosity | "<peer> tried this after Series B" | If audience is skeptical of direct claims |
| 4 | Direct | "15 min Thursday re: demand gen?" | If they're warm or you have signal-based timing |
| 5 | Pain | "Q3 demand gen plan question" | Conservative; fits exec inboxes |
3. Spam-trigger lint
After generating, lint each subject against
references/spam-trigger-words.md. Flag any that hit known triggers.
4. Critique mode
If user pastes a subject for critique, return:
Subject: "<original>"
Issues:
- <issue 1>
- <issue 2>
Better alternatives:
- "<alt 1>" — <why>
- "<alt 2>" — <why>
References
