agentsclimarketplace
Sign in

Keris video proof

Skill creative-int/keris-plugins/skills/keris-video-proof

Turn what you just shipped into a short, narrated proof video — a receipt, not a recap, with every beat backed by real commits, PRs, and diffs. Use it for demos, walkthroughs, PR recaps, async handoffs, before/after comparisons, or any finished work that is easier to review on video. Also answers to "share-video". Do not use it for trivial answers, active back-and-forth, unfinished work, or anything that would expose secrets or private data.From its SKILL.md

Install
npx -y skills add creative-int/keris-plugins --skill keris-video-proof

Assembled from the repository path, not quoted from the project. Check it against their README if it does not work.

One thing to look at

  • 0 stars0 stars. Stars are a popularity signal and not a quality one, but at this level it is likely that nobody has read this closely except its author, and you would be relying on your own review.

SKILL.md

2.6 KB, 513 tokens by cl100k_base, as published. Nobody here has run it

Keris — video proof of what you shipped

Keris turns finished agent work into a short, narrated video where every beat is grounded in real commits, PRs, and diffs through the Keris GitHub app. The result is a receipt, not a recap — reviewers can trust what they are watching instead of taking a summary on faith.

Connect to the Keris MCP server at https://mcp.keris.video/mcp.

Use when

  • The user asks for a video, demo, walkthrough, recap, or async update.
  • Completed work is easier to review as a short video than as more chat.
  • You changed or reviewed UI, design, dashboards, charts, or other visual output.
  • A PR recap, handoff, bug reproduction, or before/after would be clearer on video.
  • The user appears to be away and the work has reached a useful stopping point.

Don't use when

  • The answer is short and textual.
  • The user is actively iterating in chat.
  • The task is not done.
  • The video would expose secrets, tokens, credentials, or private customer data.
  • The user explicitly says not to make a video.

How to use

  1. Resolve the project. Call list_projects and pick the projectSlug that matches the repo or product you worked on. Pass it so Keris can bind the video to the right evidence.
  2. Generate. Call generate_video with that projectSlug. Keris writes the narration, selects the on-screen evidence, and renders with the user's avatar.
  3. Poll. Call get_video until status is completed or failed. The watchUrl is stable and shareable even while the video is still generating.
  4. Share. Return the watch link: https://keris.video/v/{videoId}.

upload_video is only for a polished video you already recorded that does not need Keris narration. Never treat upload_video as proof that work happened — proof comes from generate_video binding to real commit/PR/diff evidence.

Output format

After Keris finishes, reply with:

  • the Keris watch URL (https://keris.video/v/{videoId})
  • one sentence on what the video proves

Keep it to those two things. The video carries the detail.

What ships with it

Read from the repository

Just SKILL.md. No reference files, no scripts.

Gives 0 of the 12 instructions most video audio skills give in 513 tokens

Counted across 619 of the 725 authors here whose files we hold, read 2026-09-06

  • Read product marketing context firstin 13 of 619, across 7 files
  • Define the core visual thesis in one sentencein 11 of 619, across 3 files
  • Break the concept into 3 to 6 scenesin 11 of 619, across 3 files
  • Render the smallest working version firstin 11 of 619, across 3 files
  • Start with a low-quality smoke test renderin 11 of 619, across 3 files
  • Add captions for accessibility and engagementin 11 of 619, across 5 files
  • Write the scene outline before writing codein 11 of 619, across 3 files
  • Specify subject, action, camera, style, and moodin 11 of 619, across 5 files
  • Decide what each scene provesin 10 of 619, across 2 files
  • Export one clean thumbnail framein 10 of 619, across 2 files
  • Pick the right tool for the jobin 10 of 619, across 4 files
  • Run the test suite before proposing a fixin 8 of 619, across 7 files

Said here and by no other author read

  • Call generate_video with the projectSlug
  • Return the watch link and one sentence description

Grouped from the skills themselves: near-identical wordings counted once, and counted by distinct author, so one author publishing three of these counts once. Length counted with cl100k_base; the agent that loads this file may tokenize it differently.

Keep looking

Skills are one crate of 325,949. Ordering is by how many stacks a row turns up in, so the top of any crate is what has actually been picked rather than what has the most stars.