agentleFS
oldwinteroldwinter/skills/change-evidence/SKILL.md

change-evidence

Skill11 starsChanged 14 days agoMITView on GitHub

Capture and deliver screen-recorded evidence for UI/UX, frontend, visual, responsive, accessibility, interaction, or animation changes and their bug fixes, covering the baseline or reproduction, changed behavior, and acceptance results.

The file

We split this skill into the rules it gives an agent

3 rules. About 0 are common, 3 are distinctive, and 0 are very specific to this project.

  • ≈0

    Common · estimate

    Many files like this have a rule of this kind.

  • ≈3

    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:

    Use this skill for UI/UX, frontend, visual, responsive, accessibility, interaction, or animation changes and for bug fixes in those …

    line 10
  2. Distinctive:

    If a display, recorder, permission, or usable acceptance surface is unavailable, report that exact blocker and keep the task incomplete.

    line 20
  3. Distinctive:

    Run the smallest relevant checks after the final edit.

    line 28

≈ “Common” is an estimate: at least 1 in 5 of the 364,980 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

Change Evidence

Trigger

Use this skill for UI/UX, frontend, visual, responsive, accessibility, interaction, or animation changes and for bug fixes in those user-facing surfaces. Pure backend, data, CLI, infrastructure, documentation, and non-visual refactoring work is outside this skill unless it also changes user-facing behavior.

Completion contract

A task in scope is complete only when the handoff contains:

  • a readable recording of the baseline or bug reproduction, with the affected surface and acceptance state visible without zooming or cropping;
  • a focused recording of the changed or fixed behavior; and
  • the acceptance checks and their results, with paths to the recordings.

If a display, recorder, permission, or usable acceptance surface is unavailable, report that exact blocker and keep the task incomplete. A dry run or a verbal claim is not recording evidence.

Workflow

  1. Define the changed point and acceptance path before editing. Name the user-visible state and interaction. For a bug fix, reproduce and record the failure before changing code.
  2. Record a short baseline or reproduction focused on that path. Prefer a recorder already provided by the repository; otherwise use a platform recorder that produces a common video format, and for browser changes a Playwright or Chromium test with video capture counts as a platform recorder. Choose a duration long enough to show the path, usually 20–60 seconds. 🔴 CHECKPOINT (privacy): before recording, close or mask secrets, tokens, private sessions, and unrelated windows; start recording only after this check. Disable audio unless the user explicitly requests narration.
  3. Make the requested change while preserving the same entry point and test data where possible.
  4. Record the fixed behavior through the same focused path. Keep the affected browser or app state visible so the recording shows the observable acceptance signal.
  5. Run the smallest relevant checks after the final edit. Confirm that each recording exists, is non-empty, opens successfully, shows the changed surface, and matches the stated acceptance result.
  6. Deliver the evidence in the handoff: changed point, baseline or reproduction path, fixed path, recording paths, and exact checks/results. When the task already has a corresponding issue in scope, attach the recordings or accessible artifact links there; otherwise provide the evidence directly to the user. Do not create a new issue or publish elsewhere solely for this evidence.

Evidence quality

Keep each recording scoped to the changed point, include enough context to identify the page or test, and show the acceptance state rather than an idle desktop. Record every affected viewport or device class when the change is responsive or platform-specific. When recording fails, report the failure and its cause instead of substituting screenshots or an assertion. For headless or CI-only surfaces, capture the run's own video or trace artifact; when the run produces none, report that blocker instead of substituting screenshots.

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.