reference
/reference
Explain why a short video that performed well worked, as structured JSON. Extract the mechanisms (reason why), hook promise, viewer job, functional structure, and what a creator must be able to show to reuse the pattern. Use when analyzing a reference Reel, Short, or TikTok.
- Author
- lucastamoios ↗
- Repository
- skills ↗
- Licence
- MIT
- Last repository change
- 11.11.2026
Assessment and security
- Files in this package
- 2
- Scripts
- 0
- HTML comments
- 0
- URLs in SKILL.md
- 0
No signals detected. Repository changed on 11.11.2026, without references to retired models or APIs.
MIT, open: use and modify while retaining author attribution.
What it does and when to use it
Explain why a short video that performed well worked, as structured JSON. Extract the mechanisms (reason why), hook promise, viewer job, functional structure, and what a creator must be able to show to reuse the pattern. Use when analyzing a reference Reel, Short, or TikTok.
Original author description
Explain why a short video that performed well worked, as structured JSON. Extract the mechanisms (reason why), hook promise, viewer job, functional structure, and what a creator must be able to show to reuse the pattern. Use when analyzing a reference Reel, Short, or TikTok.
How it works
- Contract
- Method
SKILL.md preview
--- name: reference description: Explain why a short video that performed well worked, as structured JSON. Extract the mechanisms (reason why), hook promise, viewer job, functional structure, and what a creator must be able to show to reuse the pattern. Use when analyzing a reference Reel, Short, or TikTok. --- # Reference analysis Find the reason a video worked, not only what it contains. A script that copies the blocks of a reference and loses its mechanism fails (the expert called the result "dissertativo": an essay with no demonstration). ## Contract Follow [the marketing contract](../../protocol/contract.md). Input schema: [reference.input](../../protocol/schemas/reference.input.schema.json). Output schema: [reference.output](../../protocol/schemas/reference.output.schema.json). Output JSON only. ## Method 1. Read the transcript with timestamps, the frame descriptions, the caption, and the metrics. If `notes` from the expert exist, they outrank your inference: record them with `source: "expert_notes"`. 2. Write `hook`: the first line, what is on screen, and the promise. The promise is what the viewer is told they will get, in their own terms. If there is no promise, say so. 3. List `reason_why` mechanisms in order of weight. Each needs evidence: a quote or a timestamped visual. Typical ones: demonstration while talking, result promise up front, simplicity, identification with the viewer's situation and constraint, concrete or visceral contrast, proof (before and after), a list worth saving, a geographic hook, authority. Do not list a mechanism without evidence. 4. Score `jobs`: what the video does for the viewer. One of: make me laugh; that's so me; that's so us (I send it to someone); help me do something; make me feel or think. Weights are independent and need not sum to 1. 5. Describe `structure` by function. For each segment, say what happens and why it matters to the viewer. Do not just label blocks. 6. Fill `execution_requirements`: what a creator in another niche must have to reuse it. A workout demo needs a visible body result; a dev needs a screen, an artifact, or a before and after. 7. Write `transferable_pattern` as the mechanism without the niche, and `not_transferable` for elements tied to this creator or niche. 8. Compute `success.views_per_follower` when both numbers exist. It measures reach beyond the creator's audience, not shareability. Mention shares and saves when given. Return `needs_input` only when there is no transcript and no frames. A silent video with frames is valid.
Files in this package (2)
- SKILL.md 2 KB
- agents/openai.yaml 1 KB
Versions
| Repository | Score | Licence |
|---|---|---|
| This page · lucastamoios/skills | 100 | MIT |
| yuri-xyz/skills | 100 | MIT |
| zph/dotfiles | 100 | MIT |
Database snapshot: 13.11.2026 · package version bdec4f54a387 (SHA-256 digest of the entire package)
Downloads in the last 30 days: 0 · ranking
Report a problem with this skill: harmful content, incorrect descriptions or an author’s removal request.
Similar skills
- 100mirroring-reference-video
Use when someone shares a reference video, ad, Reel, Short or TikTok (link or file) and wants their own video to follow its structure, pacing or visual style - 100reel-analyzer
Feed Claude any viral reel (a link or a file) and it studies the whole thing — frame by frame plus the transcript — then breaks down exactly why it works: the hook, the beat structure, the pacing, and the on-screen visuals. Use when the user says "study this reel", "break down… - 100reel-teardown
Break down a short-form video (Instagram reel, TikTok, YouTube Short, ad) into what makes it work - the hook, pacing, every cut, on-screen text, script and ending - measured from the frames and transcript, never guessed. Use when someone pastes a reel or short link or a video… - 95short-form-video
Writes and makes short videos for TikTok, Reels, and Shorts (or 16:9 for YouTube, Reddit, LinkedIn): a hook-first script, captions, hashtags, a posting time, and, if wanted, the rendered video. Use whenever someone wants a TikTok, Reel, Short, launch or demo video, or video… - 100reel-panel
Send an 11-agent review panel to watch a video (reel, short, ad, film) and rate it - hook, pacing, retention, sound, visuals, story, CTA - each panelist from a different viewer or craft angle, then merge the scores into one ranked fix list. Use when the user asks to rate,… - 90tiktok-video-teardown
Break down a TikTok video or photo slideshow someone pastes - hook, scenes, on-screen text, sound, comments and how it performed against the creator's usual posts - and turn it into what to copy. Use when the user shares a TikTok link and asks why it worked, what happens in it,…