AgentStack
Browse Sign in
Browse Why AgentStack Sell Docs
Sign in
SKILL verified MIT Self-run

Dm Writer

skill-duncan-buildroom-profit-room-skills-dm-writer · by duncan-buildroom

Write one short, human DM or message to a specific person for a specific goal (start a conversation, share your offer, or follow up) — in your voice, so it opens a dialogue instead of pitch-slapping them. Trigger whenever the user says "write a DM", "help me message this person", "what do I say to", "reach out to [name]", "draft a message", "follow up with", "how do I bring up my offer to", or "I…

No reviews yet
0 installs
31 views
0.0% view→install

Install

$ agentstack add skill-duncan-buildroom-profit-room-skills-dm-writer

✓ scanned · ✓ verified, works with Claude Code, Cursor, and more.

Security review

✓ Passed

No issues found. Passed automated security review. · v0.1.0 How review works →

  • Prompt-injection patterns
  • Secret / credential exfiltration
  • Dangerous shell & filesystem operations
  • Untrusted network calls
  • Known-malicious package signatures

What it can access

  • Network access No
  • Filesystem access No
  • Shell / process execution No
  • Environment & secrets No
  • Dynamic code execution No

From automated source analysis of v0.1.0. “Used” means the capability is present in the source — more access means more to trust, not that it’s unsafe.

View the full security report →

Verified badge

Passed review? Show it. Paste this badge into your README, it links to the public security report.

AgentStack Verified badge Links to your public security report.
[![AgentStack Verified](https://agentstack.voostack.com/badges/verified.svg)](https://agentstack.voostack.com/security/report/skill-duncan-buildroom-profit-room-skills-dm-writer)

Reliability & compatibility

Security review passed
0 installs to date
no reviews yet
2mo ago

Declared compatibility

Claude CodeClaude Desktop

Compatibility is declared by the source manifest. End-to-end runtime verification is coming, see below.

Preview Execution monitoring

We're building live execution health for every listing: tool-call success rate, median latency, uptime, and last-checked timestamps, measured, not self-reported. It isn't live yet, so we don't show numbers we can't stand behind.

How agent discovery & health will work →
Are you the author of Dm Writer? Claim this listing to set pricing, connect Stripe payouts, and keep 70% of every sale.
Sign up to claim

About

DM Writer — one real message to one real person

A good DM is not a broadcast. It's a message to this person, referencing this relationship, with one clear reason for reaching out. This skill takes the context you give it and writes a short message that sounds like you texting someone you respect — not a funnel.

Use it for a single person. If you want a whole ranked list of who to contact first, run warm-outreach first, then bring individuals here.

Personalization

This skill works in your brand voice. Before running, load the brand profile at ~/.claude/brand-profile.md.

  • If it doesn't exist, run the personalize skill first (or just tell me your name, niche, offer, and audience — I'll create it).
  • If a field this skill needs is blank, I'll ask one or two quick questions and save the answers back to the profile so you only answer once.
  • Never invent facts or results about the user — use only what's in the profile, or ask.

Steps

1. Get the three things a good DM needs

Ask the user for these (accept short answers; infer the rest from the profile):

  • Who — the recipient: name, what they do, and any detail that makes them them (their work, a recent post, how you met).
  • Relationship — how well you know them and when you last spoke (close, warm-but-quiet, barely know them, cold-but-mutual-connection).
  • Goal — pick one: start a conversation (no ask yet) · share the offer (they've shown a fit) · follow up (you already messaged and want to nudge or re-open).

If the user only gives a name, ask the one or two questions that actually change the message — usually relationship and goal.

2. Match the message to the relationship

The register shifts with warmth, not with the goal:

  • Close / warm: casual, lowercase-fine, get to the point, a real question at the end.
  • Warm-but-quiet: reconnect first. Acknowledge the gap honestly ("been a minute"), lead with genuine interest in them, earn the right to mention what you're building.
  • Barely know / mutual connection: name the connection or the specific reason they came to mind up front, keep it very short, make the ask tiny.

3. Write the message — the non-pushy rules

  • Lead with them. First line is about the recipient or the shared context, never "So I have this offer."
  • One reason, one ask. No stacked links, no multi-paragraph pitch, no life story.
  • End on an easy question. The goal is a reply, not a yes. A question they can answer in one line beats a call-to-action.
  • Only name the offer if the goal is "share the offer" — and even then, frame it as "this might be relevant to you because [specific reason]," not a feature dump.
  • Honor the profile's words-to-avoid and tone. No fake urgency, no "quick question" bait, no flattery the user doesn't mean.
  • Keep it to 2–5 sentences. If it's longer than a text you'd actually send, cut it.

4. Add a one-line follow-up

Always include a single follow-up line the user can send if there's no reply after a few days. It should add value or lightness, never guilt ("just bumping this" / "did you see my message?"). Good follow-ups reference something new, or gracefully let it go ("no worries if the timing's off — door's open"). One follow-up only, then stop.

5. Save it

Append the message + follow-up to ~/dm-drafts.md under a heading with the recipient's name and date, so the user builds a personal swipe file of what they've sent. Confirm the path and remind them to swap in the one personal detail before sending.

Input → output example

Input: "Message Sam — we did a bootcamp together last year, haven't talked in months. He runs a small Shopify agency. Goal: share my offer (I build automated post-purchase email flows)."

Output (warm-but-quiet + share offer): > hey Sam! been way too long — how'd the agency end up going after the bootcamp? i actually thought of you because i've been building automated post-purchase email flows for Shopify stores lately, and i remember you had a bunch of clients who'd probably want that. worth a quick chat, or not really your thing right now?

Follow-up (if no reply in ~4 days): > no pressure either way Sam — if post-purchase flows aren't a priority right now, all good. would still love to catch up sometime.

Saved to ~/dm-drafts.md.

Notes / edge cases

  • If the goal is "share the offer" but the relationship is cold and there's no real fit signal, say so — recommend the "start a conversation" version instead. Pitching a stranger is exactly the move that makes selling feel gross and doesn't work.
  • Never fabricate a shared memory or a result. If the personal detail slot can't be filled truthfully, the message drops to a lighter register.
  • One message, one follow-up. If they want a whole sequence or a list of people, that's the warm-outreach skill — not more follow-ups to one person.

Source & license

This open-source skill is cataloged on AgentStack and links to its original source — we do not rehost the code.

Install and usage instructions live in the source repository linked above.

Reviews

No reviews yet, be the first.

Versions

  • v0.1.0 Imported from the upstream source.