More screenshots

    Works with the AI Agents you already use

    Claude CodeClaude CodeCursorCursorCodex CLICodex CLIMuseMuseDotsDots+21 more

    Prompt Lens: Video Analysis & Prompt Reverse Engineering

    1

    Analyze videos shot by shot and reverse-engineer a generation-ready prompt for every scene.

    $5

    Secure checkout via Stripe

    Money-back guarantee

    Converts to your local currency at checkout

    Security scanned
    Prompt Lens: Video Analysis & Prompt Reverse Engineering

    Prompt Lens: Video Analysis & Prompt Reverse Engineering

    Example session with this skill installed

    Break down this video shot by shot, save each video segment, and reverse-engineer its generation prompt. Organize the clips and prompts into a storyboard.

    • Read your context and instructions
    • Compiled the prompt lens

    Before pasting into Agensi's real-result field, replace this section with the full, verified result from the recorded test. The following describes the expected delivery format, not a transcript of an actual run.

    The output directory contains

    storyboard.md: The final storyboard. Every segment has a link to its saved MP4, its source start/end timestamps and duration, shot analysis, a complete standalone generation prompt, and relevant evidence limitations.

    clips/: Sequentially numbered, independently playable files such as scene-001.mp4 and scene-002.mp4. Continuous shots subdivided for analysis retain their shared shot-group identity.

    • analysis.json: The agent's scene-by-scene analysis and generation prompts.
    • manifest.json: Video metadata, detector settings, segment boundaries, and clip/frame paths.
    • frame-index.md and frames/: Supporting visual evidence used during analysis.

    Each storyboard section pairs the reference clip with a full prompt describing the subject, setting, action sequence, camera, composition, lighting, color, and style. Camera motion and audio are qualified when the evidence is insufficient. Video links work when storyboard.md stays with its clips/ folder; inline playback depends on the Markdown viewer.

    After reviewing the factual analysis, the user can keep it, request a rewritten story, or create an original remake using the selected timing and shot grammar. Newly generated media requires a separate generation request.

    storyboard.md

    Connects securely to your tools. The creator never sees your data.

    About this skill

    Prompt Lens analyzes a reference video shot by shot, automatically detects its real edits, and reverse-engineers a generation-ready prompt for every scene.

    It detects shot boundaries with Prompt Lens–tuned settings, preserves source timestamps, and saves an independently playable MP4 for every segment. The current agent analyzes the subject, setting, action, composition, camera language, lighting, color, style, and narrative role to produce a standalone generation prompt for each segment.

    The final storyboard.md pairs every saved video segment with its source time range, shot analysis, and full generation prompt. It links to the clips in the accompanying clips/ folder so you can review the reference and copy its prompt in one place. Sampled frames support the analysis in a separate frame-index.md; they do not replace the final storyboard. The workflow saves analysis.json and manifest.json for reuse, and checks that every segment has a prompt and an existing clip before rendering the final storyboard.

    After the automatic breakdown and prompt reconstruction are complete, the user can keep the factual analysis, rewrite the sequence as a new story, or create an original remake using the same timing and shot grammar.

    Core capabilities:

    • Automatic shot detection: Splits the video into real shots, preserves accurate timestamps, and distinguishes actual edits from duration-only analysis segments.
    • Saved video segments: Exports each analysis segment as a separate, playable MP4, retaining source audio when present.
    • Prompt reverse engineering: Derives a complete, generation-ready prompt from the visible evidence in every detected shot.
    • Clip-and-prompt storyboard: Delivers a storyboard.md containing each segment's video link, original timestamps, analysis, and full generation prompt, together with the corresponding clips/ folder.
    • Shot-by-shot visual analysis: Identifies subjects, environments, actions, camera behavior, composition, lighting, color, style, pacing, and transitions.
    • Script rewriting: Creates a new story while preserving only the timing, structure, or visual grammar selected by the user.
    • Structure-based remakes: Retains shot count, duration, pacing, and narrative function while replacing characters, settings, and creative elements.
    • Local-first processing: Video probing, shot detection, clip export, frame extraction, and storyboard rendering run on the user's device without relying on Prompt Lens servers.

    Prompt Lens is built for short-form creators, AI video makers, creative directors, storyboard artists, advertising teams, and anyone studying how effective videos are structured.

    The skill does not bypass authentication, DRM, memberships, regional restrictions, paywalls, or private access. Before installing software, downloading linked media, uploading content, calling a paid API, or generating video, the agent explains the action and asks for approval.

    How to use

    1. 1

      Connect your agent to the Agensi MCP

      Copy the prompt and paste it into Claude, ChatGPT, Cursor or any agent that supports MCP. It sets up the connection for you. Manual setup

    2. 2

      Sign in to Agensi

      Your agent shows a sign-in link. Sign in and click Allow.

    3. 3

      Ask for what you need

      Your agent loads this skill and uses it on your task. It always asks before any payment.

    Reviews

    No reviews yet

    Be one of the first to try it. Every listed skill goes through the trust checks below.

    Security scanned

    Checked by our 8-point scan before listing

    Fresh listing

    Recently published to Agensi

    Money-back guarantee

    Doesn't do what its page promises? You get your money back.

    Trust & safety

    Security scan

    Security scan passed 12 days ago

    Requires

    Python 3.9+
    FFmpeg
    FFprobe
    Python

    Permissions

    Terminal / Shell
    Read Files
    Write Files
    Network Access
    Browser

    Videos are processed locally by default. The skill accesses the network only when the user provides a supported public link and approves the download. It sends media or prompts to an external provider only when the user deliberately selects that provider and confirms the action.

    Listed12 days ago

    What's inside

    Frequently Asked Questions

    It supports direct media URLs and public video pages that the user's local yt-dlp installation can legally process. It does not bypass login requirements, DRM, memberships, regional restrictions, or private access.

    The default deliverables are saved clips from the original video, a final clip-and-prompt storyboard, analysis.json, and manifest.json. A rewritten script or remake plan is available on request. Generating new video begins only after the user selects a provider and model, reviews the final inputs, and confirms any cost.

    New in Prompt & Skill Engineering