build-a-skill

I want to turn a task I do often into a reusable Wideframe skill. Help me build it.

First, ask me:
- What is the one task you want to automate? Keep it to a single task, not a whole workflow.
- How would we know it worked? Name the objective, checkable result. If the task is a matter of taste, tell me, because skills only handle verifiable work.
- What details change each time that you should confirm with me rather than guess (naming, thresholds, target format, which footage)?

Then check it is possible:
- Look at the Wideframe tools you actually have and confirm the task can be done with them. If a step is not supported, tell me and suggest the closest alternative instead of pretending.

Then write the skill as a SKILL.md with:
- a short name and a one-line description of when to use it
- a "Confirm with the user first" section listing the details to pin down, so it fits my preferences instead of guessing
- numbered, concrete steps that use real Wideframe tools
- a "Done when" section with objective pass criteria

Then add the verification step:
- Do not grade your own work. When the output is ready, hand it to a separate verification subagent with its own context window. Give it the output, my confirmed choices, and the pass criteria. It returns specific feedback and a PASS or FAIL. On FAIL, fix the work and verify again with a new subagent until it passes.

Save it as a reusable skill so I can run it again by name. If what I described is really several tasks, help me split it into separate single-task skills.