subject-line-engine
Generate high-performing email subject lines from 18 formula categories derived from analysis of 2,684 Ben Settle emails. Use AFTER email bodies are written by E1 — subject lines must match the WRITTEN email, not a planned or hypothetical one. Produces multiple subject line candidates per email, scores them against 7 SL-specific criteria, and presents the top 5-10 for human selection. Targets 6-7 word sweet spot from corpus analysis. Trigger when users mention subject lines, email subjects, SL testing, open rate optimization, or writing subject lines for existing email drafts. Requires E1 email draft output.
Resources
4Install
npx skillscat add tonyflo79/ai-crush-vault/subject-line-engine Install via the SkillsCat registry.
E2 — Subject Line Engine
Pipeline Position: After E1 (Email Writer). Before E3 (Sequence Assembler). Subject lines must match WRITTEN emails, not planned ones.
PURPOSE
Generate high-performing email subject lines from 18 formula categories derived from analysis of 2,684 Ben Settle emails. Produces multiple candidates per email, scores them, and presents top 5-10 for human selection.
Success Criteria:
- Multiple subject line candidates per email
- All 18 formula categories considered
- 6-7 word sweet spot targeted
- Scored against 7 SL-specific criteria
- Top 5-10 presented for human selection
LAYER ARCHITECTURE
| Layer | Task |
|---|---|
| 0 | Input loading + email draft ingestion |
| 1 | Formula category selection + candidate generation |
| 2 | Scoring + ranking |
| 3 | Arena — subject line competitive generation |
REFERENCE FILES
For full execution specs, microskill details, and anti-degradation protocols:
SUBJECT-LINE-AGENT.md— Complete orchestration specificationSUBJECT-LINE-ANTI-DEGRADATION.md— Quality enforcement rulesARENA-LAYER.md— Arena specification
OUTPUT
Primary: Subject line candidates with scores per email
Location: outputs/[project-code]/email/
Install
npx skillscat add tonyflo79/ai-crush-vault/subject-line-engine Install via the SkillsCat registry.