spamguard-cold-email

---

What This Is

A Claude Code skill that scans cold email copy against a comprehensive deliverability ruleset — flags every violation, checks for ambiguous factual claims before rewriting, rewrites every flagged line in plain English, then runs a unified spam + clarity loop until both pass before anything reaches you.

Paste any copy: subject lines, email bodies, follow-up sequences, CTAs, opener lines, LinkedIn DMs, or any custom variable.

No setup. No memory file. Paste and scan.

---

What It Covers

SPAMGUARD COVERAGE
━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━

  350+  banned single words (sourced from Mailmeteor,
         ActiveCampaign, and deliverability testing)
         — word-by-word mechanical check: every token
           in your copy is looked up against the list
         — compound and hyphenated forms caught
           (lower-cost, no-cost, interest-free, money-back)
         — plurals of banned words are also banned
         — case-insensitive: Avoid, AVOID, avoid all flag

    7   high-risk phrase categories
         → Money & financial hype
         → Scammy / too-good-to-be-true language
         → Marketing overpromises
         → Pressure & clickbait
         → Health & pharma terms
         → Tech phishing-like phrases
         → Gambling, adult & blacklisted terms

   26   banned follow-up phrases
         — clichés that signal a template instantly

    1   intent check (before rewriting)
         — ambiguous claims about fees, outcomes,
           or process flagged before any rewrite
         — asks the clarifying question first

    1   unified audit loop
         — spam and clarity checked together each pass
         — grade 5 clarity: simplest word, one idea
           per sentence, no filler, no ambiguity
         — loop runs until both pass in the same scan

    5   formatting checks
         → ALL CAPS anywhere in subject or body
         → Em dashes (—) — the AI giveaway
         → Multiple exclamation marks
         → Excessive links
         → Promotional formatting

━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━

Works on:
  emails · subject lines · follow-up sequences
  CTAs · opener lines · LinkedIn DMs · custom variables

---

How It Works

---

Output Format

Every scan returns the same structure — clean, readable, actionable:

━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━
SPAMGUARD SCAN
━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━

Total violations: [n]

── Banned single words ──────────────────────
"[exact token in context]"
  → word: [banned word]
  → rewrite: [plain rewrite of the line]

── Banned phrases ───────────────────────────
"[exact phrase]"
  → rewrite: [plain rewrite]

── High-risk phrases ────────────────────────
"[exact phrase]"
  → category: [name]
  → rewrite: [fix]

── Formatting ───────────────────────────────
[flag] → [what to do]

━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━
CLEAN VERSION
Audit: [n] pass(es) — spam clean · clarity confirmed
━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━

[Full rewrite — spam-clean, grade-5 clarity, intent preserved, ready to send]

━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━
VERDICT
━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━

[One sentence: violations found, audit passes, send-ready confirmation]

---

How It Rewrites

SpamGuard does not swap words. It rewrites lines from the idea underneath them — and checks for clarity in the same pass.

THE PROCESS
━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━

  Before rewriting:
  → Check flagged lines for ambiguous claims
    about fees, outcomes, or process
  → Ask the clarifying question if needed
  → Never guess what the sender meant

  Each rewrite pass:
  → Spam check: every word against 300+ list
  → Clarity check (same pass):
     simplest word · one idea per sentence
     no filler · grade 5 · intent preserved
  → If either fails — rewrite and run again
  → Deliver only when both pass together

━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━

  Replace  →  promotional language   with  observational
  Replace  →  pressure               with  permission
  Replace  →  hype                   with  specificity
  Replace  →  complex words          with  everyday words

  If it sounds like an ad, a scam, or a coupon
  — rewrite it. If a ten-year-old would have to
  re-read it — simplify it.

---

---

Tips for Users

How to trigger a scan:

SpamGuard has two ways to run:

1. Automatic hook — After install, SpamGuard activates any time you ask about spam words, deliverability, trigger words, or banned words in Claude Code. Just ask naturally: "check this email for spam" or "does this copy have any trigger words" — the hook fires and the skill runs automatically.

2. Manual trigger — Type /spamguard in Claude Code, paste your copy, and run the scan directly. Use this when you want to run a scan without a question attached.

What to paste:

You can paste anything — full emails, subject lines only, follow-up sequences, CTAs, opener lines, LinkedIn DMs, or a single sentence you're not sure about. SpamGuard scans whatever you give it.

Custom variables in copy:

Variables like {first_name} or {company} are fine to leave in. SpamGuard will skip them during the word check and treat the surrounding copy normally.

If SpamGuard asks you a question before rewriting:

That is the Intent Check. It runs when a flagged line contains an ambiguous claim about fees, compensation, outcomes, or process. Answer the question so SpamGuard can rewrite accurately — it will never guess what you meant.

If the clean version still feels off:

Tell SpamGuard what you don't like. Say "the CTA still sounds too salesy" or "the opener is too bland" — it will run another pass. The unified loop reruns until both spam and clarity pass together.

Keeping your word list current:

Run curl -fsSL https://raw.githubusercontent.com/termsheetinator/spamguard-cold-email/main/install.sh | bash to pull the latest version any time. The word list is actively maintained — new words get added as deliverability testing surfaces new triggers.

---

Install

curl -fsSL https://raw.githubusercontent.com/termsheetinator/spamguard-cold-email/main/install.sh | bash

Open Claude Code in any directory and run:

/spamguard

Paste your copy. Get a clean scan.

Update

Already installed? Re-run the same command. It overwrites the skill and the word list with the latest version.

curl -fsSL https://raw.githubusercontent.com/termsheetinator/spamguard-cold-email/main/install.sh | bash

---

Files

File	Purpose
spamguard.md	The Claude Code skill — installs to ~/.claude/skills/
spamwords.md	Master spam word and phrase reference — installed to your project directory
install.sh	One-command installer

---

Requirements

• Claude Code — the CLI for Claude
• An Anthropic account

---