agentleFS
Faaz17Faaz17/Agents-and-Skills/skills/outreach/humanize-email/SKILL.md

humanize-email

Skill1 starsChanged 2 months agoMITView on GitHub

Strip AI writing tells from an email draft and cut it to length while preserving every fact, name, number and link exactly. Use on every draft before QA, and on any email written outside the pipeline — recruiter replies, thank-you notes, interview follow-ups.

The file

We split this skill into the rules it gives an agent

4 rules. About 0 are common, 4 are distinctive, and 0 are very specific to this project.

  • ≈0

    Common · estimate

    Many files like this have a rule of this kind.

  • ≈4

    Distinctive · estimate

    This author's own choices. Worth a look.

  • 0

    Very specific

    Tied to their project's names, files or tools. Probably skip.

  • Common
  • Distinctive
  • Very specific
  1. Distinctive:

    No contractions anywhere. Add two or three

    line 26
  2. Distinctive:

    Never add a fact, never strengthen a claim, never make an achievement sound bigger.

    line 60
  3. Distinctive:

    Never change a name, a number, a date, a URL or the signature.

    line 61
  4. Distinctive:

    If the draft's problem is the argument rather than the prose, say so and send it back.

    line 63

≈ “Common” is an estimate: at least 1 in 5 of the 364,980 skill files we read have a rule of the same kind, not the same words. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.

Keep what fits and drop the rest. Your version, ready for your agents.
Files
SKILL.md

/humanize-email, the prose pass

Usage: /humanize-email "<Person Name>/drafts/2026-08-25-space42.md"

Applies the Humanizer method (https://github.com/blader/humanizer), which the user's global instructions require as the final prose pass on anything user-facing. Cold emails are the highest-stakes case: a recipient who thinks an email was machine-written stops reading, and the whole research effort is wasted.

Steps

  1. Read the draft and count the body words.
  2. Cut first. Length reduction removes most tells for free. Over 150 words: delete a whole paragraph, not a bit from each. The paragraph to delete is almost always the one listing a second project or explaining why internships are valuable.
  3. Run the tell sweep. Full list in rules/email_format.md. The high-frequency offenders:
    • More than one em dash. Replace with full stops
    • I hope this email finds you well · I am reaching out to · I am writing to express my interest
    • passionate about · leverage · utilize · delve · robust · seamless · cutting-edge · at the intersection of
    • Rule-of-three lists that are not genuinely three things
    • It's not just X, it's Y
    • Uniform sentence length. This is the subtlest and most reliable tell. Break the rhythm with one short sentence
    • A tidy summarising sentence before the sign-off. Delete it — the ask should be the last thing
    • No contractions anywhere. Add two or three
  4. Fix the ask. One ask, answerable in one line, in the final paragraph, with nothing after it but the sign-off.
  5. Run the swap test on paragraph 1. Substitute a competitor's name. If it still reads fine, the opening has no research in it and the draft goes back to email_writer.
  6. Verify facts survived exactly. Numbers, dates, award names, publication venues, technology names, URLs, both names, the signature block. Check each against profile.md. List them in the report — this check is why the skill can be trusted to rewrite freely everywhere else.
  7. Set status: edited in the frontmatter.

Output format

## Humanized: <file>

Words: <before> → <after>

**Cut**
- <what and why>

**Tells fixed**
| Was | Now |
|---|---|

**Ask:** <as it now reads>
**Swap test:** pass | fail
**Facts verified unchanged**
- <fact> ✓

**Verdict:** ready for send_qa | back to email_writer — <reason>

Rules

  • Never add a fact, never strengthen a claim, never make an achievement sound bigger.
  • Never change a name, a number, a date, a URL or the signature.
  • Shorter wins ties.
  • If the draft's problem is the argument rather than the prose, say so and send it back. Polishing a wrong angle produces a well-written email nobody answers.
  • This skill does not clear a draft for sending. send_qa does that.

Download now (ZIP)

Does it work?

Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.

No reviews yet

People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.

  • It worked
  • Partly
  • It didn't
What did you use it for? e.g. Reviewing pull requests in a Rails app

Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.

Your trail

Files you open will collect here, so you can find them again.

Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.

Now do this for your own.

Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.