SightHouse helps reverse engineers understand unknown binaries by matching their functions against a database of known signatures. When it finds a match, it tells you the original function name, its prototype, and which library or framework it came from.

Unlike hand-curated signature sets, SightHouse keeps its database up to date automatically: it continuously scrapes, compiles, and analyzes open-source or proprietary projects, and then lets you match any binary against it from your favorite reverse-engineering tool.

How it works

SightHouse is split into three components, each serving a different need:

• SRE Clients: plugins for IDA Pro, Ghidra, and Binary Ninja that query a SightHouse server from inside your disassembler and annotate matched functions.
• Frontend: a self-hostable API server that answers signature queries from the clients, backed by a signature database.
• Signature Pipeline: the automated engine that grows the database: it scrapes open-source projects (e.g. from PlatformIO), compiles them, and extracts function signatures using Ghidra BSIM.

Which part do I need?

Most users only need the SRE Clients. Not sure? Check the docs quarkslab.github.io/sighthouse.

Installation

SightHouse is available on PyPI. Install only the component you need:

pip install sighthouse-client     # SRE clients (IDA, Ghidra, Binary Ninja)
pip install sighthouse-frontend   # Frontend API server
pip install sighthouse-pipeline   # Signature pipeline
pip install sighthouse[all]       # Everything

Installing an SRE client plugin

After installing sighthouse-client, register the plugin for your disassembler:

# IDA
sighthouse client install ida --ida-dir /path/to/ida_dir

# Ghidra (requires Ghidra 11.3+ with pyGhidra)
sighthouse client install ghidra --ghidra-install-dir /path/to/ghidra

# Binary Ninja
sighthouse client install binja

Ghidra note: the plugin runs on pyGhidra, so launch Ghidra with $GHIDRA_INSTALL_DIR/support/pyghidraRun rather than the classic launcher.

See the client installation guide for per-tool details and install-from-source options.

From source

git clone https://github.com/quarkslab/sighthouse && cd sighthouse
# Creates a virtual env and installs SightHouse into it
make install

Docker

Pre-built images are published for deploying a pipeline, a frontend, or a headless Ghidra:

docker pull ghcr.io/quarkslab/sighthouse/sighthouse:latest
docker pull ghcr.io/quarkslab/sighthouse/sighthouse-pipeline:latest
docker pull ghcr.io/quarkslab/sighthouse/sighthouse-frontend:latest

Usage

Once a client plugin is installed and pointed at a SightHouse server, open a binary in your disassembler and run the plugin:

• IDA: go to Edit -> Plugins -> SightHouse Client Plugin, enter your credentials and the server endpoint. Matches are added as comments.
• Ghidra: open a program and run SightHouseClientGhidra.py from the Script Manager, fill in the prompt, and press Submit. Matches appear on the corresponding functions.
• Binary Ninja: set your username, password, and endpoint under Edit -> Settings (search for "SightHouse"), then run it from the Plugins menu. Matches are added as tags.

Full walkthroughs with screenshots are in the client quickstart.

Documentation

Full documentation is available at quarkslab.github.io/sighthouse, including:

• Client installation and quickstart
• Frontend installation and quickstart
• Signature pipeline, how it works, and writing custom modules
• API reference

Building the docs locally

git clone https://github.com/quarkslab/sighthouse && cd sighthouse/doc
make install   # install documentation dependencies
make serve     # serve the docs at a local URL

Running the tests

git clone https://github.com/quarkslab/sighthouse && cd sighthouse
make check     # run the format test
make test      # run the unit test

Changelog

• 1.0.6: Add the SightHouse web panel; add server-side Function ID support; revamp BobRoss; improve documentation.
• 1.0.5: Migrate object storage from MinIO to RustFS; improve test suite.
• 1.0.4: Improve Docker Compose setup and deployment scripts.
• 1.0.3: Update Docker images.
• 1.0.2: Improve the client installer.
• 1.0.1: Improve documentation.
• 1.0.0: First public release of SightHouse.

Contributing

Contributions, issues, and feature requests are welcome. If you hit a bug or want to see a new format or SRE tool supported, please open an issue.

Authors

• MadSquirrels (Forgette Benoit)
• Fenrisfulsur (Babigeon Sami)

License

SightHouse is licensed under the MIT License.