agentleFS
xmartlabsxmartlabs/agent-skills/skills/pr-evidence/SKILL.md

pr-evidence

Skill2 starsChanged 18 days agoMITView on GitHub

Capture visual evidence for a pull request that touches UI or email — a screenshot for a static UI change, a GIF (plus a linked mp4) for an interactive flow, or a rendered-email screenshot for an email change — upload it as native GitHub attachments (or use a configured legacy host), and post it as a PR comment (optionally linking it from an issue tracker). Best-effort and non-blocking. Use after a PR is created when the diff touches UI or email.

4 other projects use a version of this.

Adopted by

Most recently taken up by

  • Other versions4 projects use a version of this.

    Word for word · 0

    None found.

    Edited · 0

    None found.

    Wrote their own version · 4

    1. heristop/leclap★ 35same name
    2. ArefMozafari/pr-evidence★ 19same name
    3. aiatelie/ai-atelie★ 9same name
    4. acgs-ai/ACGS★ 2same 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 repository2 other files it gives its agents.

The file

We split this skill into the rules it gives an agent

7 rules. 0 are common, 0 are distinctive, and 7 are very specific to this project.

  • 0

    Common

    Most of the 3 edited versions we read keep these.

  • 0

    Distinctive

    This author's own choices. Worth a look.

  • 7

    Very specific

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

  • Common
  • Distinctive
  • Very specific
  1. Very specific:

    Follow flow/pr-evidence-flow.md — that document is the single source of truth for every stage of this flow (guard → detect kind → boot → …

    line 9
  2. Very specific:

    Do not duplicate its content here.

    line 11
  3. Very specific:

    All paths in this skill (flow/, scripts/, config/) are relative to the directory this SKILL.md lives in — the skill is self-contained and …

    line 14
  4. Very specific:

    When invoked without explicit inputs (e.g.

    line 26
  5. Very specific:

    KIND: leave unset unless the user explicitly named one.

    line 38
  6. Very specific:

    Run scripts/run.sh check first (flow Stage 0).

    line 42
  7. Very specific:

    After the guard passes, verify the flow prerequisites, including updating GitHub CLI to gh >= 2.99.0, before booting.

    line 45

“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

PR Evidence

Follow flow/pr-evidence-flow.md — that document is the single source of truth for every stage of this flow (guard → detect kind → boot → authenticate → capture → host → comment → cleanup). Do not duplicate its content here; read it before acting. Mobile targets add flow/mobile-evidence-flow.md on top.

All paths in this skill (flow/, scripts/, config/) are relative to the directory this SKILL.md lives in — the skill is self-contained and installs the same way in every harness.

Inputs

  • TICKET — the issue/ticket key this PR closes, if any (used to namespace hosted artifacts).
  • PR_NUMBER — the pull request number to comment on.
  • CHANGED — the changed-file list for the branch, used to detect kind and infer route/role in Stage 4.
  • KIND — optional override for the detected kind (screenshot, recording, email); when unset, derive it from CHANGED per flow Stage 1.

Resolving them for the current PR

When invoked without explicit inputs (e.g. /pr-evidence), resolve them from the current branch:

  1. PR_NUMBER: gh pr view --json number --jq '.number'
  2. TICKET: a ticket-key-shaped token (e.g. PC-123, or this project's convention) in the branch name (git branch --show-current), falling back to the PR title (gh pr view --json title --jq '.title'). If neither carries one, proceed without — hosting still works, it just won't namespace by ticket.
  3. CHANGED:
    base="$(gh pr view --json baseRefName --jq '.baseRefName')"
    git diff "origin/$base...HEAD" --name-only
  4. KIND: leave unset unless the user explicitly named one.

Before anything else

Run scripts/run.sh check first (flow Stage 0). Treat a non-zero exit as an unconditional stop — do not boot, authenticate, or capture against an origin that didn't pass this check.

After the guard passes, verify the flow prerequisites, including updating GitHub CLI to gh >= 2.99.0, before booting. Native attachments use Stage 6 directly; see the flow for details.

Contract

This flow is best-effort and non-blocking: any stage failing means log a warning, skip the remaining stages, and report "evidence pending — generate manually" — never abort or fail the caller's PR flow.

Optional: firing automatically on PR creation

Some projects wire their own PR-creation step to invoke this skill right after a PR opens. That wiring is opt-in and lives in the consuming project, not in this kit. If you do wire it, one hard rule from the flow doc's trust-model section applies: read .pr-evidence/config.json from the repository's base branch, never from the PR's head branch. Its boot, auth.script and host.upload values are executed as code; reading them from an untrusted PR head would let a fork PR inject an arbitrary command that runs the moment the PR is opened.

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.