Top AI Repos — open-source AI, indexed and scored
Top AI Repos tracks AI repositories on GitHub and answers two different questions about each one: is it moving right now, and would you bet a product on it.
Top AI Repos tracks AI repositories on GitHub and answers two different questions about each one: is it moving right now, and would you bet a product on it.
A minimal LLM-powered zero-day vulnerability scanner by AISLE.
| Date | Stars |
|---|---|
| 2026-07-31 | 302 |
| 2026-08-11 | 305 |
| 2026-08-18 | 308 |
| 2026-08-20 | 310 |
| 2026-08-21 | 311 |
| 2026-08-23 | 310 |
| 2026-08-26 | 313 |
| 2026-08-27 | 318 |
| 2026-08-30 | 319 |
| 2026-09-01 | 320 |
| 2026-09-03 | 333 |
| 2026-09-04 | 339 |
| 2026-09-05 | 342 |
| 2026-09-06 | 342 |
| 2026-09-07 | 343 |
| 2026-09-08 | 344 |
| 2026-09-09 | 346 |
| 2026-09-10 | 348 |
| 2026-09-11 | 349 |
| 2026-09-13 | 352 |
| 2026-09-16 | 353 |
| 2026-09-17 | 355 |
| 2026-09-18 | 357 |
| 2026-09-20 | 358 |
| 2026-09-23 | 357 |
| 2026-09-25 | 359 |
| 2026-09-29 | 360 |
| 2026-09-30 | 360 |
| 2026-10-03 | 359 |
| 2026-10-04 | 362 |
| 2026-10-07 | 364 |
| 2026-10-09 | 365 |
| 2026-10-10 | 366 |
| 2026-10-15 | 366 |
| 2026-10-17 | 367 |
| 2026-10-18 | 368 |
| 2026-10-21 | 368 |
| 2026-10-23 | 369 |
| 2026-10-31 | 371 |
| 2026-11-01 | 372 |
| 2026-11-04 | 372 |
Today
— stars today
This week
+3 stars this week
This month
+10 stars this month
Momentum
0.0
growth rate 0.81%/day
# Nano-analyzer **A minimal LLM-powered zero-day vulnerability scanner by [AISLE](https://aisle.com).**  > **Research prototype for demonstration purposes.** This is a simple, single-file harness that is able to detect real zero-day vulnerabilities. Note that it is a prototype, biased towards C/C++ memory safety bugs, and will produce false positives. We are sharing it as-is in the spirit of open research — expect sharp corners. ## What it does Nano-analyzer is a simple single-file Python scanner that sends source code through a three-stage LLM pipeline: 1. **Context generation** — a model writes a security briefing about the file: what it does, where untrusted data flows, which buffers exist and how big they are. 2. **Vulnerability scan** — the same model, primed with the context, hunts for zero-day bugs function by function and outputs structured findings. 3. **Skeptical triage** — each finding is challenged over multiple rounds by a skeptical reviewer that can grep the codebase to verify (or refute) defenses. An arbiter makes the final call. Results are saved as Markdown and JSON files for human review. ## Current limitations This is a v0.1 prototype. Please keep the following in mind: - **C/C++ bias.** The prompts, few-shot examples, and heuristics are heavily tuned for C/C++ memory safety vulnerabilities (buffer overflows, NULL derefs, integer overflows, type confusion). It will scan other languages but is much less effective there. - **False positives.** Even with multi-round triage, expect findings that don't hold up on closer inspection. Always verify manually. - **False negatives.** The scanner can miss entire vulnerability classes — logic bugs, race conditions, cryptographic issues, authentication bypasses, etc. A clean scan does not mean the code is safe. - **Single-file analysis.** Each file is scanned independently. Cross-file vulnerabilities that depend on interactions between compilation units will likely be missed. - **LLM-dependent.** Results vary with the model used. Different models will find different things and hallucinate different false positives. ## Setup ### Requirements - Python 3.8+ - An OpenAI API key (for OpenAI models) or an OpenRouter API key (for other providers) - Optional: [ripgrep](https://github.com/BurntSushi/ripgrep) (`rg`) for triage grep lookups - Optional: [Google codesearch](https://github.com/google/codesearch) (`csearch`/`cindex`) for faster grep on large repos ### Install ```bash git clone https://github.com/weareaisle/nano-analyzer.git cd nano-analyzer # No dependency installation needed. Run directly: python3 scan.py --help ``` ### API keys Set your API key as an environment variable: ```bash # For OpenAI models (model names without a slash, e.g. "gpt-5.4-nano"): export OPENAI_API_KEY=sk-... # For OpenRouter models (model names with a slash, e.g. "qwen/qwen3-32b"): export OPENROUTER_API_KEY=sk-or-... ``` The scanner determines which key to use based on the model name: if it contains a `/`, it routes through OpenRouter; otherwise it uses the OpenAI API directly. ## Usage ### Basic scan ```bash # Scan a single file python3 scan.py ./path/to/file.c # Scan a directory recursively python3 scan.py ./path/to/src/ ``` ### Common options ```bash # Use a different model python3 scan.py ./src --model gpt-5.4 # Control parallelism python3 scan.py ./src --parallel 30 # Point triage grep at the full repo root (useful when scanning a subdirectory) python3 scan.py ./lib/crypto/ --repo-dir ./ # Only surface high-confidence findings python3 scan.py ./src --min-confidence 0.7 # More triage rounds for higher accuracy (default: 5) python3 scan.py ./src --triage-rounds 7 ``` ### All flags | Flag | Default | Description | |------|---------|-------------| | `path` | *(required)* | File or directory to scan | | `--model` | `gpt-5.4-nano` | Model for all stages (context, scan, triage) | | `--parallel` | `50` | Max concur
Excerpt of 6,413 characters
Read on GitHubWould you bet a product on this? Bounded 0–100 and slow moving.
matched fp:66caad6a81e498ad, llm:Description: 'A minimal LLM-powered zero-day vulnerability scanner by AISLE.' Language: Python. Purpose: vulnerability scanner using LLM.
matched fp:66caad6a81e498ad, llm:Description: 'A minimal LLM-powered zero-day vulnerability scanner by AISLE.' Language: Python. Purpose: vulnerability scanner using LLM.
matched fp:66caad6a81e498ad, llm:Description: 'A minimal LLM-powered zero-day vulnerability scanner by AISLE.' Language: Python. Purpose: vulnerability scanner using LLM.