agentleFS
dtran320dtran320/coderegon-trail/skills/narration/SKILL.md

Narration

Skill1 starsChanged 7 months agoMITView on GitHub

This skill should be used when generating spoken explanations of code, diffs, or architecture. Provides guidelines for conversational, TTS-friendly narration that sounds like a staff engineer pairing with a new team member.

3 other projects use a version of this.

Adopted by

Most recently taken up by

  • Other versions3 projects use a version of this.

    Word for word · 0

    None found.

    Edited · 0

    None found.

    Wrote their own version · 3

    1. leventilo/mobius★ 8same name
    2. bmohan111/claude-skills★ 0same name
    3. TimeCraker/vibe-studio★ 0same name

    How matches are found. Word for word: the exact same text. Edited: shares passages with this file (“99% same” means 99% of its five-word phrases also appear here; 10% or more counts). Their own version: similar meaning, or, weakest, the same name. We can’t tell which project had it first. Projects are counted among the 200 most-starred candidates, and mentions are links to the repository we found, up to 50 per site, so both counts are a floor.

  • In this repository3 other files it gives its agents.

The file

We split this skill into the rules it gives an agent

11 rules. About 0 are common, 10 are distinctive, and 1 is 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.

  • 1

    Very specific

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

  • Common
  • Distinctive
  • Very specific
  1. Distinctive:

    Avoid reading code syntax literally.

    line 22
  2. Distinctive:

    Use natural descriptions: "a function that validates user input" not "function validateUserInput parenthesis data colon string"

    line 23
  3. Distinctive:

    Keep sentences under 25 words for natural speech rhythm

    line 24
  4. Distinctive:

    Use punctuation to create pauses — commas, periods, em dashes

    line 25
  5. Distinctive:

    Avoid parenthetical asides (they break speech flow)

    line 26
  6. Very specific:

    ElevenLabs only: Use <break time="1.0s" /> tags to insert pauses between sections or before important points.

    line 27
  7. Distinctive:

    Don't read code line by line: "Line 1 import express from express, line 2..."

    line 66
  8. Distinctive:

    Don't use jargon without context: "It implements memoization via closure capture"

    line 67
  9. Distinctive:

    Don't be vague: "Some changes were made to improve things"

    line 68
  10. Distinctive:

    Don't narrate boilerplate: skip import statements, config lines, formatting changes

    line 69
  11. Distinctive:

    Don't repeat yourself between overview and detail modes

    line 70

≈ “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

Narration Skill

Generate spoken narration for code walkthroughs and diff reviews. All narration will be piped to TTS, so it must sound natural when read aloud.

Core Principles

  1. Conversational — Talk like a helpful colleague, not a textbook. Use "Let's look at...", "Notice how...", "The reason for this is..."
  2. Educational — Explain WHY changes were made or WHY code is structured a way, not just WHAT the code does
  3. Progressive — Start high-level, let the user drill into detail on demand
  4. Engaging — Vary pacing, use transitions ("Now the interesting part is..."), highlight non-obvious decisions

TTS-Friendly Rules

  • Spell out abbreviations on first use: "JWT, or JSON Web Token"
  • Say "slash" for path separators: "server slash routes slash users dot ts"
  • Avoid reading code syntax literally. Never say "const x equals await..."
  • Use natural descriptions: "a function that validates user input" not "function validateUserInput parenthesis data colon string"
  • Keep sentences under 25 words for natural speech rhythm
  • Use punctuation to create pauses — commas, periods, em dashes
  • Avoid parenthetical asides (they break speech flow)
  • ElevenLabs only: Use <break time="1.0s" /> tags to insert pauses between sections or before important points. Do NOT use break tags with say or openai engines — they will be read literally

Narration Templates

Section Overview (Diff Review)

Let's look at [section title]. This is [critical/important/supporting].
The changes here [describe purpose in plain language].
[If breaking:] This is a breaking change because [reason].
There are [N] files involved: [list files with "slash" separators].
[If cross-cutting:] This impacts [areas of the codebase].

Section Overview (Walkthrough)

Now we're looking at [area of codebase]. This is where [purpose].
The key thing to understand is [core concept].
You'll see [pattern] used here because [reason].
The main files are [list naturally].

Detailed Explanation

Looking at [filename]:
[For each significant change or section:]
- [Describe intent] by [describe approach]
- [Highlight key decisions or trade-offs]
- [Note edge cases or error handling]

Transitions Between Sections

"Now let's move on to [next section]. This connects to what we just saw because..."
"That covers the [previous topic]. Next up is [next topic], which is where..."
"Good, so we understand [summary]. The next piece of the puzzle is..."

What NOT to Do

  • Don't read code line by line: "Line 1 import express from express, line 2..."
  • Don't use jargon without context: "It implements memoization via closure capture"
  • Don't be vague: "Some changes were made to improve things"
  • Don't narrate boilerplate: skip import statements, config lines, formatting changes
  • Don't repeat yourself between overview and detail modes

Verbosity Levels

Brief

2-3 sentences per section. Just the what and why. Skip how.

Normal (default)

5-8 sentences per section. What, why, and key how. Mention patterns and decisions.

Detailed

Full walkthrough. What, why, how, trade-offs, alternatives considered, edge cases. 10+ sentences.

Fly-Through Narration

Fly-Through Opening

Let's fly through the key code in this section. I'll show you N snippets, walking through the execution flow. Watch the code as I narrate.

Fly-Through Step

[2-4 sentences per snippet. Reference visible code directly: "See that try-catch on line 23?"]
[End with transition to next: "This calls into the service layer, which we'll see next."]

Fly-Through Closing

That covers the key code paths for this section. We looked at N files, from [starting point] through to [ending point]. The main takeaway is [key insight].

Quiz Narration

Quiz Introduction

Let's check your understanding of what we've covered. I'll ask N quick questions about the last few sections. Just answer in your own words — no need to be exact.

Quiz Question

[Read the question slowly and clearly]
[Pause to let the user think]

Correct Answer Feedback

That's right. [1-2 sentences elaborating, referencing the code]. Look at [file:line] to see exactly where this happens.

Partial Answer Feedback

You're on the right track. [Acknowledge what's correct]. The piece you're missing is [key concept]. Check [file:line] for the full picture.

Incorrect Answer Feedback

Not quite. [Gentle explanation of the correct answer]. If you look back at [file:line], you can see [specific evidence]. The key thing is [main takeaway].

Quiz Summary

Nice work — you got N out of M. [If high score: "You're tracking well."] [If low: "Let's keep going — these concepts will click as we see more code."] Your running total is X out of Y.

References

  • references/voice-personas.md — Voice assignments and persona styles for team mode
  • references/ranking.md — Importance ranking criteria for ordering sections
  • references/fly-through.md — Fly-through snippet plan format and display template
  • references/quiz.md — Quiz question types, evaluation guidelines, and narration style

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.