agentleFS
agiea-brainagiea-brain/agency-skills/skills/tapajyoti-das/win-to-angles/SKILL.md

win-to-angles

Skill1 starsChanged 17 days agoMITView on GitHub

Use this skill when something real happened with a client and it is sitting unused because the story feels too specific to publish or too confidential to tell. Produces three to five distinct angles out of the one story, each with a drafted opening, who it is aimed at, and what has to be removed or approved first. Also use for "is there a post in this", "turn this into content", "can I write about this client", "we had a win".

The file

We split this skill into the rules it gives an agent

10 rules. About 0 are common, 10 are distinctive, and 0 are very specific to this project.

  • ≈0

    Common · estimate

    Many files like this have a rule of this kind.

  • ≈10

    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:

    Run the confidentiality pass on each angle: what identifies the client, whether the story survives without it, what needs approval.

    line 37
  2. Distinctive:

    MUST work from the record of what happened, not only the owner's recollection.

    line 51
  3. Distinctive:

    MUST identify what was surprising, or say the story is not content.

    line 52
  4. Distinctive:

    MUST produce genuinely distinct angles, not variations of one.

    line 53
  5. Distinctive:

    MUST map each angle to a theme the firm already talks about.

    line 54
  6. Distinctive:

    MUST run a confidentiality pass on every angle.

    line 55
  7. Distinctive:

    MUST flag angles that are recognisable even anonymised.

    line 56
  8. Distinctive:

    NEVER mark a story publish-ready when the client has not been confirmed as nameable.

    line 57
  9. Distinctive:

    NEVER invent a detail to make the story land better.

    line 58
  10. Distinctive:

    NEVER draw a lesson the story does not support.

    line 59

≈ “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 to one specific thing that happened, not to a content calendar. Produces several separate angles from it, with the confidentiality work done.

The play

  1. Get the sequence of what actually happened from the record, including what the client said at the time. The owner's summary a month later has already smoothed out the part that made it interesting.
  2. Find what was genuinely surprising, and this is the whole exercise. Not the outcome, which is usually what the firm sells and therefore predictable, but the thing the firm did not expect. If nothing was surprising, say so: the story is a case study, not content, and forcing it into a post produces the kind of writing people scroll past.
  3. Pull three to five separate angles, not one point reworded. Different angles come from different places in the story:
    • The counterintuitive lesson, where the obvious move was wrong.
    • The thing the firm got wrong first. The strongest angle available and the one owners resist, because it is also the one that reads as honest rather than promotional.
    • The pattern across clients that this one confirmed.
    • The process detail nobody in the category talks about.
    • The client's own framing of their problem, which is often sharper than the firm's.
  4. Map each angle to something the firm already talks about. An angle that fits none of its themes either expands them on purpose or gets dropped, because a one-off post in an unfamiliar register confuses the people who follow the firm for something else.
  5. Draft the opening line or two for each angle in the firm's voice. Enough for the owner to tell whether they would actually post it, which is the only test that matters at this stage.
  6. Rank by who the firm is trying to reach right now, not by which angle is most interesting to the firm. Those are different and the second one is a trap.
  7. Run the confidentiality pass on each angle: what identifies the client, whether the story survives without it, what needs approval. Most stories work with the industry and the size kept and everything else dropped.
  8. Flag any angle that needs permission even anonymised, because anyone in that client's market would recognise the engagement. Getting this wrong costs a client to win a post.

What good looks like

  • First thing I check that most people skip: whether anything in the story actually surprised us. If the outcome was exactly what we sell, there is no post in it, and forcing one produces the writing people scroll past.
  • The mediocre version: the win announcement. A number, a logo, a thank-you. It performs for the team and for nobody else.
  • I know it is right when one of the angles is something we got wrong first. That is the one that gets replies, because it is the one that costs something to publish.
  • Second test: the angles are different arguments, not one point in five registers.
  • Red flag: wanting to post it before anybody has asked the client. One post is never worth a client, and the version that feels most postable is usually the most identifiable.

Rules

  • MUST work from the record of what happened, not only the owner's recollection.
  • MUST identify what was surprising, or say the story is not content.
  • MUST produce genuinely distinct angles, not variations of one.
  • MUST map each angle to a theme the firm already talks about.
  • MUST run a confidentiality pass on every angle.
  • MUST flag angles that are recognisable even anonymised.
  • NEVER mark a story publish-ready when the client has not been confirmed as nameable.
  • NEVER invent a detail to make the story land better.
  • NEVER draw a lesson the story does not support.

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.