Spam Word Checker Claude Skill
The Spam Word Checker Claude skill applies a fixed set of spam and deliverability rules to cold email subject lines, openers, follow-ups and CTAs. It flags banned words and phrases and suggests plainer rewrites. It is free and open source under the MIT license, from the GrowthEngineX cold outbound skills repo, and it runs without any tools or keys.
git clone https://github.com/growthenginenowoslawski/coldoutboundskills.git ~/cold-email-ai-skillscd ~/cold-email-ai-skillsnpx skills add growthenginenowoslawski/coldoutboundskills --skill spam-word-checkerWhat the Spam Word Checker skill does
Spam filters and deliverability checkers can penalize wording that sounds like an ad or a scam. This skill turns a list of lessons into rules Claude applies to every line of cold copy it writes or reviews.
The lists cover banned single words, short phrases, promotional and pressure wording, phishing-style language and categories that do not belong in B2B copy. It also bans em dashes, all caps, greetings before the first name and third-person company references.
It adds rules for opt-out lines: a line should not promise to stop after silence, since the sequence keeps going. It then supplies safer replacement patterns and ways to handle a banned word inside a company name.
- Checks subject lines, openers, follow-ups and CTAs against banned word and phrase lists
- Catches close variants, so hyphens and splits do not hide a banned root word
- Flags pressure, promotional and phishing-style wording
- Enforces formatting rules: no em dashes, all caps, multiple exclamation marks or greeting prefixes
- Rewrites closeout lines so they ask for an explicit reply instead of promising to stop on silence
- Offers safe replacement phrases for common risky lines
- Suggests renamed forms of company names that contain banned words
How it works
Scan the copy
It checks each subject line, opener, follow-up and CTA against the banned words, phrases and style rules.
Catch variants
It treats close variants as banned when the root word is still clear, even with hyphens or splits.
Rewrite risky lines
It replaces hype with plain, observational language and pressure with permission, using its safe replacement table.
Fix closeouts and names
It rewrites silence-based opt-out lines and adjusts company names that contain banned words.
Example prompts
Once the Spam Word Checker skill is installed, ask Claude in plain language. These prompts trigger it:
- Check this cold email for spam words
- Is this subject line safe to send?
- QA my follow-up sequence for deliverability risks
- Scan this opener and rewrite anything that sounds like an ad
- Fix the unsubscribe line in this email so it doesn't promise to stop on silence
What you get
- A list of flagged words, phrases and style issues
- A rewritten version of the copy
- Safer replacement wording for each flagged line
- A cleaned closeout line
Who it is for
- Cold email writers who want a consistent QA pass
- Agencies checking client copy before launch
- Anyone using Claude to generate outbound copy at volume
How to install the Spam Word Checker skill
What you need: Nothing beyond Claude. Works with Claude Code and Claude.ai.
- Clone
git clone https://github.com/growthenginenowoslawski/coldoutboundskills.git ~/cold-email-ai-skills - Open in Claude Code
cd ~/cold-email-ai-skills - Skills CLI
npx skills add growthenginenowoslawski/coldoutboundskills --skill spam-word-checker - Manual
mkdir -p ~/.claude/skills && cp -r ~/cold-email-ai-skills/skills/spam-word-checker ~/.claude/skills/spam-word-checker
After installing, restart Claude Code or start a new session. Claude loads the Spam Word Checker skill automatically when your request matches it, or you can ask for it by name.
Tips for better results
- Run it on the final draft, then re-read it aloud, since rewrites can sound stiff.
- Add any word your own deliverability tool flags to the banned list. The skill says to treat those as banned going forward.
- Use it together with the cold-email-copywriting skill, which it is designed to back up.
Limitations
- It is a fixed word list from the author's experience. It does not test real inbox placement.
- The list is strict, and some banned words are common in normal English, so you may need approved exceptions.
Spam Word Checker Claude skill FAQ
Yes. It is MIT licensed and published in the growthenginenowoslawski/coldoutboundskills repo on GitHub. It is a text-only skill and needs no API key or paid tool.
Clone the repo and open it in Claude Code. You can also run npx skills add growthenginenowoslawski/coldoutboundskills --skill spam-word-checker, or copy the spam-word-checker folder into ~/.claude/skills.
It works in both. It is a single SKILL.md of rules with no scripts, so you can zip the folder and upload it in Claude.ai under Settings, Capabilities, Skills on a paid plan.
This skill bans lists of single words, pressure phrases like act now, promotional claims and phishing-style language. It also bans em dashes and all caps. The full lists are in the SKILL.md.
No. It reduces risky wording, but placement also depends on domain setup, list quality and sending habits. Pair it with a deliverability audit to check the infrastructure side.