agentleFS
hranesshraness/personal-monorepo-template/.agents/skills/riff/SKILL.md

riff

Skill4 starsChanged 49 days agoMITView on GitHub

Turn a rambly spoken or stream-of-consciousness transcript into a clean, faithful first-person knowledge-base note under kb/riffs/. Use when the user pastes a dictated voice dump, says “riff” or “add a riff,” or asks to clean up a journal-style transcription. Preserve every useful idea, aside, hedge, attribution, and coined term; execute embedded transcriber research requests; use attachments as context without embedding the raw files.

3 other projects use a version of this.

Adopted by

Most recently taken up by

The file

We split this skill into the rules it gives an agent

8 rules. 0 are common, 6 are distinctive, and 2 are very specific to this project.

  • 0

    Common

    Most of the 3 edited versions we read keep these.

  • 6

    Distinctive

    This author's own choices. Worth a look.

  • 2

    Very specific

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

  • Common
  • Distinctive
  • Very specific
  1. Distinctive:

    Write the result to:

    line 14
  2. Very specific:

    Create a finished, self-contained note that reads as if the speaker wrote it deliberately.

    line 20
  3. Distinctive:

    Keep every useful idea, aside, uncertainty, hedge, attribution, and coined term.

    line 25
  4. Distinctive:

    Keep the note in first person and preserve the speaker's vocabulary.

    line 26
  5. Distinctive:

    Treat “sidebar for the transcriber,” “look this up,” “flesh this out,” and similar phrases as instructions, not prose to preserve.

    line 32
  6. Very specific:

    Put substantial companion research in an appropriately named kb/ note and link it once from the riff.

    line 35
  7. Distinctive:

    Use attached screenshots, whiteboards, or documents to verify names and recover meaning.

    line 41
  8. Distinctive:

    Read the entire transcript before writing, complete embedded follow-ups, write the note and any companion files, then report the paths …

    line 43

“Common” is measured: kept by at least 6 in 10 of the 3 edited versions of this file we could read. “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

Clean up a riff

Write the result to:

kb/riffs/<YYYY-MM-DD>-<descriptive-slug>.md

Create a finished, self-contained note that reads as if the speaker wrote it deliberately. Follow WRITING.md: lead with the point, prefer active voice, name concrete details, and cut filler.

Preserve the person

  • Repair run-ons, punctuation, dropped words, and obvious transcription mistakes.
  • Keep every useful idea, aside, uncertainty, hedge, attribution, and coined term.
  • Keep the note in first person and preserve the speaker's vocabulary.
  • Reorder scattered threads into a coherent arc when needed. Use headings only when the note genuinely spans multiple topics.
  • End on the strongest natural beat. Do not manufacture a conclusion or add a recap.

Resolve transcriber instructions

Treat “sidebar for the transcriber,” “look this up,” “flesh this out,” and similar phrases as instructions, not prose to preserve.

  • Perform the requested lookup or research and integrate the result into the finished note.
  • Put substantial companion research in an appropriately named kb/ note and link it once from the riff.
  • State what remains open only when the follow-up cannot be completed.
  • Attribute additions only when a reader could otherwise mistake them for the speaker's claim.

An optional editorial note must be brief, clearly labeled, and genuinely additive. Omit it when it only restates the riff.

Use attached screenshots, whiteboards, or documents to verify names and recover meaning. Describe relevant content in prose; do not copy or embed the raw attachment in the riff.

Read the entire transcript before writing, complete embedded follow-ups, write the note and any companion files, then report the paths created and any material inference that remains uncertain.

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.