Official GGUF builds of WardenSight-1B, converted by the model author from the original safetensors (SHA-256 070345c0…11f16) with llama.cpp's converter.

WardenSight-1B is a 1-billion parameter terminal agent specialized for vulnerability file localization in WordPress plugins: given a plugin repository and a CWE class, it explores the code through a read-only terminal interface and submits a ranked list of files it believes contain that weakness. It is a triage instrument — it tells an analyst which files to read first; it does not produce line numbers, exploits, patches, or proof that code is vulnerable.

See the main model card for the full evaluation (File-F1 0.3263 vs base 0.2671 on 1,646 held-out WordPress plugin tasks, +22.2% relative), intended use, limitations, and the agent prompt contract.

Runs anywhere llama.cpp does, including Apple Metal:

Or load directly in LM Studio / Ollama. WardenSight is an agentic model: it expects the terminal-agent prompt contract described in the main model card and performs best driven by the Antares CLI harness documented there.

Apache-2.0, inherited unchanged from the safetensors release.