agentleFS
Sign inStart free
shankerdhandshankerdhand/claude-second-brain/.claude/skills/draft-letter/SKILL.md

draft-letter

Skill0 starsChanged 2 months agoMITView on GitHub

Draft a letter or formal email in the user's own voice from a template. Use when the user invokes /draft-letter or asks to write/draft a letter, application, complaint, family letter, or formal email.

The file

We split this skill into the rules it gives an agent

8 rules. About 0 are common, 3 are distinctive, and 5 are very specific to this project.

  • ≈0

    Common · estimate

    Many files like this have a rule of this kind.

  • ≈3

    Distinctive · estimate

    This author's own choices. Worth a look.

  • 5

    Very specific

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

  • Common
  • Distinctive
  • Very specific
  1. Very specific:

    Write it so it sounds like {{OWNER_NAME}}, not like an assistant.

    line 6
  2. Very specific:

    Read brain/identity/voice.md.

    line 10
  3. Very specific:

    Read brain/writing/snippets.md for signature and boilerplate.

    line 11
  4. Very specific:

    If it's to a known person, read their file in brain/people/ — context is what makes a letter land.

    line 12
  5. Distinctive:

    Ask in one message, not one at a time:

    line 19
  6. Distinctive:

    If the brain already answers one of these, don't ask it.

    line 25
  7. Distinctive:

    Follow the template's own "Rules" section — those exist for a reason.

    line 29
  8. Very specific:

    Never invent facts: no amounts, no dates, no events, no family news.

    line 30

≈ “Common” is an estimate: at least 1 in 5 of the 350,495 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

Write it so it sounds like {{OWNER_NAME}}, not like an assistant.

Before writing — always

  1. Read brain/identity/voice.md. This governs everything below it.
  2. Read brain/writing/snippets.md for signature and boilerplate.
  3. If it's to a known person, read their file in brain/people/ — context is what makes a letter land.
  4. Pick a template from brain/writing/letters/ or brain/writing/emails/. If none fits, write from scratch and follow the same discipline.

Ask only what you can't infer

Ask in one message, not one at a time:

  • Who is it to, and what's your relationship?
  • What's the one outcome you want from it?
  • Any facts, dates, or reference numbers that must appear?
  • Anything that must not be mentioned?

If the brain already answers one of these, don't ask it.

Draft

  • Follow the template's own "Rules" section — those exist for a reason.
  • Never invent facts: no amounts, no dates, no events, no family news. If a slot has no supplied fact, leave it as <…> and list what's missing under the draft.
  • Match the language the relationship is actually conducted in — if letters home are in another language or a mix, write it that way.
  • Formal letters: subject line carries the reference number, one ask, a date, a contact.

Present

Show the full draft, then:

Missing: <the facts you need to fill the brackets>
Tone check: <what you aimed for> — say if you want it warmer / blunter / shorter.

Rules

  • Never send. Draft only. If {{OWNER_NAME}} wants it sent, create an email draft and say it's waiting — sending needs an explicit yes on the final text.
  • Offer a second version at a different register if the right tone is genuinely unclear.
  • Once approved, save it to brain/writing/letters/sent/ or emails/sent/ with the date and recipient in the filename. These become voice samples.

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.

Check your setup, free