| name | cold-email-audit |
| description | 30-point audit of a B2B outbound program across four dimensions — infrastructure (8 points), targeting (8 points), messaging (8 points), and operations (6 points). Produces a 0-100 score, the top 3 levers, and a 90-day remediation order. Loaded by the main cold-email skill when the user asks to audit or grade their outbound. Based on the JMC 30-Point Outbound Audit framework. |
| user-invocable | false |
| allowed-tools | ["Read","Write","Grep","WebFetch"] |
Cold Email Audit — 30-Point Outbound Program Review
Loaded by cold-email when the user asks to "audit our outbound",
"why isn't this working", "grade our cold email program", or similar.
Scores a B2B outbound program against 30 points across 4 dimensions
and surfaces the top 3 levers.
Activation triggers
- "Audit our outbound"
- "Why isn't our cold email working"
- "Grade our outbound program"
- "Outbound audit"
- "What's wrong with our cold email"
- "Score our outbound"
Workflow
1. Intake (one batch of questions)
Ask the user to provide context in a single message — DON'T ask 30
questions sequentially. Surface a structured intake form:
To run the 30-point audit, share what you can across these 4 dimensions:
INFRASTRUCTURE
- Sending domain(s) — primary and secondary
- Inbox provider (GSuite / O365 / Smartlead / Instantly / custom)
- Warm-up: how many mailboxes, ramped how long, current daily limits
- SPF / DKIM / DMARC status (or paste DMARC TXT record)
TARGETING
- ICP definition (one paragraph)
- PSP (Pain Signal Profile) — what signals do you anchor on?
- List source (apollo / clay / hand-built / scraped / referrals)
- List size, send volume per week, response rate baseline
MESSAGING
- Paste your current T1 opener
- Paste your subject line (or top 3 subjects)
- How many sequence touches, cadence in days
- Personalization layer: none / variable / per-person-research
OPERATIONS
- Reply handling: who, response SLA
- Booking motion: link in email, link in reply, manual?
- Hand-off to sales: who routes, when, what info
- Reporting cadence: weekly / monthly / never
Don't grade until you have the user's answers. If they don't have
a dimension at all, that's a finding — note it as "Not in place".
2. Score against the 30-point rubric
Load references/audit-rubric.md and score each point 0 / 0.5 / 1.
The four dimensions:
| Dimension | Points | What it covers |
|---|
| Infrastructure | 8 | Domains, warm-up, DNS, sending architecture |
| Targeting | 8 | ICP, PSP, list source, segmentation |
| Messaging | 8 | Opener, body, CTA, sequence, personalization |
| Operations | 6 | Reply SLA, booking motion, hand-off, reporting |
Total: 30 points → normalize to a 0-100 score:
Score = (totalPoints / 30) * 100
3. Identify the top 3 levers
After scoring, find the 3 lowest-scoring points and rank them by
impact-per-day-of-effort. Don't just list the lowest scores — weight
by:
- Compounding impact: a fix to infrastructure compounds across
every campaign forever; a subject-line tweak compounds for one
campaign.
- Effort: an SPF flatten is 1 hour; a full PSP rewrite is 2 weeks.
- Sequencing: deliverability MUST be solid before any messaging
iteration matters.
4. Output
Produce a structured report:
# Outbound Audit — <company name>
## Score: <0-100>/100
| Dimension | Score | Grade |
|---|---|---|
| Infrastructure | <0-8>/8 | A / B / C / D / F |
| Targeting | <0-8>/8 | |
| Messaging | <0-8>/8 | |
| Operations | <0-6>/6 | |
## The top 3 levers
### 1. <Lever name> — <effort / impact>
<2-3 sentences on what to fix, why it compounds, how to ship in 7-14 days.>
### 2. <Lever name>
...
### 3. <Lever name>
...
## 90-day remediation order
| Days | Lever | Owner | Verify by |
|---|---|---|---|
| 1-14 | <Lever 1> | <role> | <metric> |
| 15-45 | <Lever 2> | | |
| 46-90 | <Lever 3> | | |
## Full point-by-point detail
<Optional: only if user asks. Otherwise keep concise.>
Grade thresholds
| Score | Grade | Meaning |
|---|
| 90-100 | A | Mature program. Tune for scale. |
| 75-89 | B | Solid foundation. Iterate on messaging + ops. |
| 60-74 | C | Works in spurts. Fix the bottom-third before scaling. |
| 40-59 | D | Don't scale spend until top-3 levers are fixed. |
| 0-39 | F | Stop sending. Rebuild infrastructure first. |
References
references/audit-rubric.md — full 30-point rubric with scoring criteria
../../cold-email/references/jmc-framework.md — the framework underneath
- The full 30-Point Outbound Audit Template ships as a DOCX in
Operator Pass