CtrlK
BlogDocsLog inGet started
Tessl Logo

gamussa/reels-producer-skill

help quicky produce instagram reels and youtube shorts

94

2.12x
Quality

96%

Does it follow best practices?

Impact

87%

2.12x

Average score across 3 eval scenarios

SecuritybySnyk

Passed

No findings from the security scan

Overview
Quality
Evals
Security
Files

task.mdevals/scenario-3/

Reel Project Kickoff: Training Session Footage

Problem Description

A fitness coach is ready to produce a 30-second highlight reel from a recent training session. The footage — a mix of short training clips and one recording from a camera that was accidentally left running for the entire event — is in a footage/ directory representing media on an external drive. Before any editing can happen, the project environment needs to be confirmed, the workspace set up, and a thorough inventory run on the footage so the coach knows exactly what they're working with.

The coach suspects there may be a problem clip in the bunch and wants a clear written summary of what the probe finds before committing to any editing decisions. They also want the project organized so that future sessions can resume without losing track of where things stand.

Before starting: run footage/prepare.sh to generate the sample clips for this session. Once that completes, the footage files will be in footage/ and you can proceed with setup.

Output Specification

Set up the project environment and run an inventory of the footage, then produce the following:

  1. A scaffolded project directory using the reel-building pipeline tools (the scripts/ directory contains the relevant tooling), targeting a 30-second reel.

  2. Connect the footage from footage/ into the project the way the pipeline recommends for media that lives on an external drive. Note any important caveat the coach should be aware of regarding the footage source.

  3. work/clips.json — run the pipeline's clip inventory tool against the footage, making sure it is configured for a 30-second target reel.

  4. work/probe_summary.md — a written summary for the coach. It must cover: how the working environment and workspace were prepared for this session (what was verified and how the footage was connected), total number of clips, combined runtime, any clips that are problematic or need special handling before the next step, and any clips that appear to use a log or HDR color profile.

  5. Handle any clips that can't go straight into the pipeline — apply whatever the pipeline recommends for clips that aren't yet ready for normalization. Outputs should be placed where the pipeline expects them.

Clean up any large temporary files you produce along the way; the footage/ directory itself is fine to leave as-is.

.mcp.json

tile.json