agentleFS
rosslevinskyrosslevinsky/portable-agent-skills/skills/demo-video/SKILL.md

demo-video

Skill12 starsChanged 45 days agoApache-2.0View on GitHub

On-demand guided-tour walkthrough video of a built feature — the last skill in the v2 workflow. Drives an existing Playwright / frontend setup through the primary flow slowly with pauses, records video, and derives timed subtitles from as-built.md and test-step timing; optional background music. Never bootstraps Playwright and never claims success from "a video exists" alone. Without ffmpeg it still ships Playwright's native video plus a sidecar subtitle file; it degrades to ordered screenshots only when no video can be recorded at all. Narration audio is out of scope — it writes subtitles, not speech. Use when the user invokes /demo-video, or says "make a demo video", "record a walkthrough", "generate a guided tour of the feature".

27 other projects use a version of this.

Adopted by

Most recently taken up by

Used with

Projects with it also use karpathy-coder, ai-act-readiness and 10 more.

Other files people use alongside this one.

Files found in the same public projects as a version of this one, more often than they turn up anywhere else. Each project counts once, large collections count for less, and this file's own repository is left out.
  • csiddhant796-blipSkill

    karpathy-coder

    Use when writing, reviewing, or committing code to enforce Karpathy's 4 coding principles — surface assumptions before coding, keep it simple, make surgical changes, define verifiable goals. Triggers on "review my diff", "check complexity", "am I overcomplicating this", "karpathy check", "before I commit", or any code quality concern where the LLM might be overcoding.

    Also in 4 projects that use this

  • Hussain-ElfalahSkill

    ai-act-readiness

    Also in 3 projects that use this

  • Hussain-ElfalahSkill

    c-level-agents

    Also in 3 projects that use this

  • csiddhant796-blipSkill

    self-eval

    Honestly evaluate AI work quality using a two-axis scoring system. Use after completing a task, code review, or work session to get an unbiased assessment. Detects score inflation, forces devil's advocate reasoning, and persists scores across sessions.

    Also in 4 projects that use this

  • +8 more from Hussain-Elfalah

The file

We split this skill into the rules it gives an agent

3 rules. 0 are common, 2 are distinctive, and 1 is very specific to this project.

  • 0

    Common

    Most of the 25 edited versions we read keep these.

  • 2

    Distinctive

    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:

    Write a Playwright guided-tour spec that walks the primary flow slowly: one clear action per step, an explicit wait/pause long enough to …

    line 61
  2. Distinctive:

    If ffmpeg is available, optionally extract change-point frames, encode/re-encode the final video, mux in the subtitles, and add optional …

    line 75
  3. Very specific:

    All output — video, extracted frames, subtitles, music — is heavy and disposable.

    line 83

“Common” is measured: kept by at least 6 in 10 of the 25 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

Demo Video — Guided-Tour Walkthrough

Classification: Degraded — the guided-tour spec and its subtitle/narration script are produced in any runtime; recording the walkthrough video needs a Playwright/frontend setup, and muxing captions, adding music, and extracting change-point frames need ffmpeg. Without ffmpeg the skill still delivers Playwright's native .webm plus a sidecar subtitle file; without Playwright video it degrades to ordered screenshots + the narration script text. It is only Runtime-limited if no walkthrough artifact can be produced at all.

Overview

Produce a guided-tour walkthrough video of a feature that has already been built: a slow, paced Playwright tour of the primary flow, recorded to video, with subtitles derived from as-built.md and the tour's step timing, and optional background music.

This is a standalone, on-demand skill and the last piece of the v2 workflow. It reuses the same video infrastructure as the web-verify skill (if that skill is unavailable, drive the existing Playwright setup directly) and uses as-built.md as the narration source.

Two rules:

  • Never bootstrap tooling. It uses the repo's existing Playwright/frontend setup and degrades when it is absent — it does not install anything.
  • A produced file is not success. The tour must actually show the feature working; when video can't be rendered, the honest fallback is ordered screenshots + the narration script, not a claim that "a video exists."

TTS/voice narration is out of scope: this skill produces subtitles, not audio.

Step 1 — Check prerequisites (never bootstrap)

Without installing anything, check for:

  • An existing Playwright / frontend setup (as the web-verify skill detects). If it is unavailable there is no driver to capture frames with, so the recorded tour is not merely skipped — nothing can produce one. Degrade to a narration script, plus a storyboard built from stills the user already has or captures by hand, and say which half is missing and why. Do not describe the fallback as "ordered screenshots" without saying where they come from: the tool that would have taken them is the one absent.
  • ffmpeg, for frame extraction, re-encoding, subtitle muxing, and music. If absent, still keep Playwright's natively recorded .webm and ship it with a sidecar subtitle file (.vtt/.srt) — only the muxed-in captions, music, and frame extraction are skipped (Degraded). Fall back to ordered screenshots + a subtitle/script file only when Playwright video recording itself is unavailable.
  • as-built.md as the narration source. If it is missing, ask the user for a short tour outline; if operating autonomously, derive the outline from the plan's success criteria and note the assumption.

Step 2 — Author the guided-tour spec

Write a Playwright guided-tour spec that walks the primary flow slowly: one clear action per step, an explicit wait/pause long enough to read at each step, and a highlight of the element in focus. See references/guided-tour-spec.md for the pattern (slow-motion config, per-step pauses, stable anchors).

Step 3 — Record and derive subtitles

Record the tour to video. Derive one caption per step from as-built.md (what the step demonstrates) timed to that step's start and duration, and write a standard subtitle file (.vtt or .srt). See references/subtitles.md for the timing derivation and the optional-music note.

Step 4 — Render (optional, heavier layer)

If ffmpeg is available, optionally extract change-point frames, encode/re-encode the final video, mux in the subtitles, and add optional background music. If ffmpeg is absent, deliver Playwright's natively recorded .webm alongside the sidecar subtitle file from Step 3 (captions ride as a separate file rather than muxed in); fall back to the screenshots + subtitle/script text only when no video was recorded at all.

Step 5 — Artifact retention

All output — video, extracted frames, subtitles, music — is heavy and disposable. Ensure the repo gitignores it, scoped to the output directory (demo-video-output/, including any frames subdirectory under it) — not bare repo-wide globs like *.webm/*.vtt/*.srt, which can silently untrack committed media or caption assets elsewhere in the repo. Reference the result by path or CI URL only; never commit it.

Reporting

Report: the flow(s) toured, whether a rendered video was produced or the skill degraded to screenshots + script, the subtitle/script path, and the artifact paths/URLs.

References

  • references/guided-tour-spec.md — the guided-tour Playwright spec pattern (slow-motion, per-step pauses, highlights).
  • references/subtitles.md — deriving timed subtitles from as-built.md + step timing, the optional-music note, and why narration audio is out of scope.

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.