agentleFS
DevinoSolutionsDevinoSolutions/claude-plugins/skills/editing-demos/SKILL.md

editing-demos

Skill0 starsChanged 9 days agoMITView on GitHub

Edits finished Demofy demos and their look: reads a demo's script beat by beat, proposes changes to narration, cuts, camera, look and voiceover without spending anything, renders the approved edit as a new demo, records a demo again when the page changed, and reads or updates the workspace brand kit (default look, brand colour, voice, closing call to action). Use when the user asks 'change the narration in step 3', 'cut the pricing part', 'make the captions bigger', 'turn off the voiceover', 're-record this demo, the page changed', 'use our brand colour on new demos', or 'which voice do my demos use'. Rendering uses the plan's allowance, so it always confirms first.

The file

We split this skill into the rules it gives an agent

8 rules. About 0 are common, 1 is distinctive, and 7 are very specific to this project.

  • ≈0

    Common · estimate

    Many files like this have a rule of this kind.

  • ≈1

    Distinctive · estimate

    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:

    Every tool in this skill comes from the demofy MCP server, the Demofy connector, and is written here as demofy:<tool>.

    line 13
  2. Very specific:

    Tools listed by name: call them directly, for example demofy:demofy_get_script.

    line 15
  3. Very specific:

    When the user wants it rendered, call demofy:demofy_ship_edit with the jobId and the edit exactly as proposed.

    line 39
  4. Very specific:

    Ask which the user means: the same steps on a changed page (demofy:demofy_rerender) or a fresh plan from the original brief …

    line 45
  5. Very specific:

    Call the tool with the jobId.

    line 46
  6. Very specific:

    Call demofy:demofy_brand_kit.

    line 51
  7. Distinctive:

    Never render, re-record or refresh without the user's explicit yes to a summary that names the demo and says it uses the allowance.

    line 57
  8. Very specific:

    Never include a confirmToken the user has not approved, and never reuse one for different arguments.

    line 59

≈ “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
<!-- Generated from plugins/demofy/skills/editing-demos/SKILL.md by scripts/sync-skills.mjs. Edit the plugin copy, then run the script. -->

Demofy: editing demos

Change what a finished demo says and shows, or the look new demos start from, with the user approving every render.

Connecting and calling the tools

Every tool in this skill comes from the demofy MCP server, the Demofy connector, and is written here as demofy:<tool>.

  • Tools listed by name: call them directly, for example demofy:demofy_get_script.
  • Only demofy:search_tools and demofy:execute_typescript listed (Code Mode, the default): call demofy:search_tools first, then call each tool inside demofy:execute_typescript as external_<tool> (for example external_demofy_get_script); the program must return its result. Validation and confirmation are identical on both surfaces.
  • "Tool not found": switch to the Code Mode route instead of retrying the direct call.
  • No demofy tools at all: Demofy is not connected yet. Follow the setup-demofy skill.

A refused call throws an Error whose message starts with its code, for example CONFLICT:, QUOTA_EXCEEDED: or CONFIRMATION_REQUIRED:.

Tools you will use

  • demofy:demofy_get_script: a demo as the editor sees it: every beat numbered from 1 with its title, narration, whether it is cut, and its camera choices; the look and the names each look field accepts; the voiceover switch; and whether it is editable.
  • demofy:demofy_propose_edit: turns a list of changes into an edit WITHOUT rendering: what would change, what was skipped and why, the edit itself, and its path (restyle re-styles the kept recording, refilm films the page again). Nothing is spent or saved. Change kinds: look {field, value}, say {beat, text} (empty text makes a beat silent), trim {beat, cut}, camera {beat, choice}, voiceover {value}.
  • demofy:demofy_ship_edit: renders the proposed edit, unchanged, as a NEW demo; the original stays. Uses the allowance. Two-step confirmation.
  • demofy:demofy_rerender: records a finished demo again from its stored plan, as a new demo, for when the page changed or the recording went wrong. Two-step confirmation.
  • demofy:demofy_refresh: makes a demo again from its original page and brief, planned afresh, for when the product changed too much for the old steps. Two-step confirmation.
  • demofy:demofy_brand_kit and demofy:save_brand_kit: read and save the workspace's default look, brand colour (Pro), narration voice and closing call to action. Saving replaces the whole kit; owners and admins only.
  • demofy:demofy_list_voices and demofy:preview_voice: narration voices, and a short audio sample of one.

Workflow

Editing one demo

  1. Identify exactly one demo (use the managing-demos skill to find it) and call demofy:demofy_get_script. If editable is false, say why (not finished, or uploaded) and stop.
  2. Map the user's request to changes using the beat numbers and look names from the script. Call demofy:demofy_propose_edit with them.
  3. Show the user the changes, anything skipped and why, and the path in plain words: a restyle keeps the recording, a refilm records the page again. Adjust and propose again until the user is happy with it.
  4. When the user wants it rendered, call demofy:demofy_ship_edit with the jobId and the edit exactly as proposed. The first call does not render: it returns CONFIRMATION_REQUIRED with a confirmToken. End that program there.
  5. Tell the user it renders a new demo in a few minutes, uses one demo from the allowance, and leaves the original as it is. Wait for an explicit yes, then call demofy:demofy_ship_edit again with the SAME arguments plus the confirmToken, in a new call.
  6. Give the new demo's page, https://app.demofy.io/jobs/<id>, and follow it with demofy:demofy_demo_status as in the making-demo-videos skill.

Recording again

  1. Ask which the user means: the same steps on a changed page (demofy:demofy_rerender) or a fresh plan from the original brief (demofy:demofy_refresh).
  2. Call the tool with the jobId. It returns CONFIRMATION_REQUIRED with a confirmToken. End that program.
  3. Tell the user it records a new demo, keeps the original, and uses one demo from the allowance. Wait for an explicit yes, then call the same tool again with the SAME arguments plus the confirmToken.

Changing the brand kit

  1. Call demofy:demofy_brand_kit. If editable is false, say only an owner or admin can change it.
  2. Change only the fields the user asked about and keep every other field as read; use voice ids from demofy:demofy_list_voices (offer demofy:preview_voice if they want to hear one). A brand colour needs the Pro plan; colorState says whether it applies.
  3. Show the before and after of the changed fields, wait for a yes, then call demofy:save_brand_kit with the whole kit. It applies to new demos, not to finished ones.

Rules

  • Never render, re-record or refresh without the user's explicit yes to a summary that names the demo and says it uses the allowance.
  • Ship only the edit that demofy:demofy_propose_edit returned, unchanged. Never write an edit by hand.
  • Never include a confirmToken the user has not approved, and never reuse one for different arguments; it works once and lasts 10 minutes.
  • On QUOTA_EXCEEDED, say the allowance is used up and offer the plan options (demofy:list_plans) instead of retrying.

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.