agentleFS
agiea-brainagiea-brain/agency-skills/skills/andrew-kappel/morning-paper-outreach/SKILL.md

morning-paper-outreach

Skill1 starsChanged 17 days agoMITView on GitHub

Use this skill early each morning to turn a curated publication you subscribe to into warm, specific touchpoints. Produces a short list of the day's stories worth sending, who each one goes to and why, and a drafted note for each that you review before sending. Also use for "who should I send this to", "daily reading for client outreach", "turn the paper into touchpoints", "curate and cultivate".

The file

We split this skill into the rules it gives an agent

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

  • ≈0

    Common · estimate

    Many files like this have a rule of this kind.

  • ≈8

    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:

    MUST read the source through your own paid subscription, and never circumvent a paywall.

    line 43
  2. Distinctive:

    MUST link to the article rather than pasting its text, and keep any quoted line short.

    line 44
  3. Distinctive:

    MUST have a specific reason per pairing, naming something about that person.

    line 45
  4. Distinctive:

    MUST cap the day at five, and send fewer on a thin day.

    line 46
  5. Distinctive:

    MUST review every note before sending.

    line 47
  6. Distinctive:

    MUST check last-contacted before every send.

    line 48
  7. Distinctive:

    NEVER attach an ask, a meeting request or a pitch to one of these notes.

    line 49
  8. Distinctive:

    NEVER send to someone who has not responded to the last two.

    line 50

≈ “Common” is an estimate: at least 1 in 5 of the 372,429 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

Applies first thing each morning, on a schedule, against a publication you pay for. Produces three to five article-and-person pairs with a drafted note each, for review before anything is sent.

The play

  1. Read the day's edition of the source you subscribe to, in full, through your own account. One strong curated source beats a feed: the value here comes from the paper being edited by someone good, so the agent is choosing among pieces that already cleared a bar. Other subscriptions can be swapped in; what cannot be swapped in is anything you do not have the right to read.
  2. Hold a people file: clients, prospects, partners, friends, with what each one works on, what they have said they care about, and when they were last contacted. The fields and how to build the file from scratch are in references/people-file.md. Without it the picks are generic, and a generic send is worse than no send.
  3. Match stories to people on a specific reason, not a topic overlap. "Logistics person, logistics article" is not a reason. The test, the four kinds of match that work, and the do-not-send list are in references/matching-rules.md.
  4. Cap the day at three to five pairs. The cap is the discipline that keeps this from becoming a mailing list. If a day's edition yields one good pair, send one.
  5. Draft a note per pair, two or three sentences, that says why it made you think of them and asks nothing. The patterns, with examples, are in references/note-patterns.md. A note with an ask attached converts the gesture into outreach and the recipient feels the difference immediately.
  6. Branch by the relationship:
    • Client, mid-engagement: only send something that bears on their work, never something that reads as marketing.
    • Prospect in an open deal: send sparingly. During a live negotiation, do not send at all.
    • Lapsed contact, over a year: acknowledge the gap in a half sentence and move on. Pretending it has not been a year is the thing people notice.
    • Friend: looser, shorter, and no business framing at all.
    • Anyone who has not replied to the last two sends: stop. The rule is in references/matching-rules.md.
  7. Review every note before it goes. The skill drafts and never sends, because the entire value is that the recipient believes a person chose this for them.
  8. Record what you sent, what you edited and what you killed. Tomorrow's picks read yesterday's record. The log format and what each signal changes are in references/feedback-loop.md.

What good looks like

  • First thing I check that most people skip: when I last contacted them. A perfect article match to someone contacted twice this month is worse than no send, and the people file is the only thing that catches it.
  • The mediocre version: five articles a day to the same twenty people, each with "thought you'd find this interesting". It is a newsletter nobody subscribed to, and it teaches the whole list to ignore you.
  • I know a pick is right when the note could not be sent to anyone else without rewriting it.
  • I know the system works when people reply about the article rather than thanking you for sending it.
  • Red flag: the same five names every week. Either the people file is too small or the matching has collapsed into topic overlap.

Rules

  • MUST read the source through your own paid subscription, and never circumvent a paywall.
  • MUST link to the article rather than pasting its text, and keep any quoted line short.
  • MUST have a specific reason per pairing, naming something about that person.
  • MUST cap the day at five, and send fewer on a thin day.
  • MUST review every note before sending. The skill never sends on its own.
  • MUST check last-contacted before every send.
  • NEVER attach an ask, a meeting request or a pitch to one of these notes.
  • NEVER send to someone who has not responded to the last two.

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.