agentleFS
GabrieleDattileGabrieleDattile/agent-superpowers/skills/reverse-engineering/blackbox-cloner/SKILL.md

blackbox-cloner

Skill1 starsChanged 15 days agoMITView on GitHub

Clean-room software reverse engineering engine. Deconstructs proprietary tools, desktop utilities, and SaaS features (e.g. Adobe tools, Figma canvas, background removers) into functional specs and synthesizes clean-room open-source implementations.

The file

We split this skill into the rules it gives an agent

4 rules. About 0 are common, 3 are distinctive, and 1 is very specific to this project.

  • ≈0

    Common · estimate

    Many files like this have a rule of this kind.

  • ≈3

    Distinctive · estimate

    This author's own choices. Worth a look.

  • 1

    Very specific

    Tied to their project's names, files or tools. Probably skip.

  • Common
  • Distinctive
  • Very specific
  1. Distinctive:

    Zero Copyright Contamination: No proprietary headers, decompiled assembly, or closed blobs are used.

    line 27
  2. Very specific:

    Auditability: Every generated scaffold includes an independent Clean-Room Specification (cleanroom-spec.md).

    line 29
  3. Distinctive:

    Run the Blackbox Cloner CLI:

    line 51
  4. Distinctive:

    Verify all keyboard shortcuts, mouse drag-and-drop, and state mutations match the intended behavior.

    line 62

≈ “Common” is an estimate: at least 1 in 5 of the 350,495 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

🕵️ Skill: Blackbox Cloner (Clean-Room Tool Replicator)

1. When to Trigger (Routing)

  • TRIGGER WHEN:
    • The user asks to reverse engineer, clone, or replicate a proprietary tool or feature (e.g., "replicate this Adobe Photoshop layer blending tool", "clone the Figma auto-layout engine", "reverse engineer this SaaS background-removal widget").
    • The user wants a clean-room open-source alternative to a closed-source software component.
    • The user provides UI screenshots, functional descriptions, or sample input/output files and wants an equivalent standalone implementation.
  • DO NOT TRIGGER WHEN:
    • The user only wants to capture network HTTP requests into an SDK (trigger api-synthesizer instead).
    • The task is simply reading an existing open-source repository's dead code (trigger arch-archaeologist instead).

2. Fundamental Philosophy & Clean-Room Doctrine

"True software replication does not copy proprietary source code or binaries. It observes black-box behaviors, formulates mathematical and state-machine abstractions, and implements an independent, clean-room solution using modern open standards."

This doctrine guarantees:

  1. Zero Copyright Contamination: No proprietary headers, decompiled assembly, or closed blobs are used.
  2. Modern Stack Upgrades: Replaces legacy desktop architectures with WebAssembly, HTML5 Canvas, WebGL, Web Workers, or modern Python libraries (NumPy, Pillow).
  3. Auditability: Every generated scaffold includes an independent Clean-Room Specification (cleanroom-spec.md).

3. Operational Protocol

Phase 1: Black-Box Behavior Extraction

Analyze the target tool along four structural dimensions:

  1. State Machine & Data Model:
    • Entities (e.g. Layers, Nodes, Canvas, Selection, History Stack).
    • Undo/Redo transactions and immutable state transitions.
  2. Event & Interaction Matrix:
    • Pointer events (drag, hover, pan, pinch-zoom, multi-touch).
    • Keyboard hotkeys and shortcuts (e.g., Spacebar for Pan, V for Selection, Ctrl+Z).
  3. Core Processing Engine:
    • Mathematical transformations (affine matrix 2D/3D, blending formulas, pixel filters).
    • Rendering strategy (Canvas 2D context, WebGL shaders, SVG DOM, or headless Python pipeline).
  4. Serialization & Formats:
    • Input ingestion formats (raw images, vector paths, JSON, binary blobs).
    • Export pipelines (PNG, SVG, PDF, JSON state dumps).

Phase 2: CLI Deconstruction & Scaffolding

Run the Blackbox Cloner CLI:

# Deconstruct and generate clean-room spec
python skills/reverse-engineering/blackbox-cloner/scripts/cloner.py deconstruct --tool "LayeredCanvasEditor" --category "graphics"

# Synthesize a ready-to-run clean-room web or python scaffold
python skills/reverse-engineering/blackbox-cloner/scripts/cloner.py scaffold --tool "LayeredCanvasEditor" --target "web" --out-dir "./scaffolds/canvas-editor"

Phase 3: Verification & Independence Testing

  1. Confirm the generated scaffold runs with zero external proprietary dependencies.
  2. Verify all keyboard shortcuts, mouse drag-and-drop, and state mutations match the intended behavior.
  3. Validate that generated code is fully typed (TypeScript or Python with Type Hints).

4. Guardrails & Negative Constraints

  • ❌ NEVER inspect or incorporate leaked, decompiled, or proprietary source code. The replication must strictly follow black-box observable behaviors.
  • ❌ NEVER produce fragile monolithic scripts. Separate State, Interaction Handling, and Processing Engines into distinct modules.
  • ❌ ALWAYS emit an immutable History Stack (undo/redo) for interactive tools.

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.