Commit graph

4 commits

Author SHA1 Message Date
trvon
9161ed2b98
eval: CTF binary scan — 3.7% binary detection vs 66.7% source
- Binary scan: 1/27 findings (only gets() import detected)
- Source scan: 18/27 findings (66.7% TP rate)
- Fuzz harnesses: generated but cov:2 (Swift ABI issue)
- Binary backend needs significant improvement
2026-05-29 20:03:07 -04:00
trvon
a36801cf8e
perf: Z3 LRU cache, batch observers, fold-rust body cache, ripgrep benchmark
- Z3 query LRU cache (4096 entries, O(1) promote/evict)
- Batch observer dispatch in security walker
- fold-rust: single-walk hash, dead code removed, timing added
- bench/corpora/rust-oss/: ripgrep benchmark harness
2026-05-27 19:06:32 -04:00
trvon
c4e42cec0a
Add scan router and MCP job queue
Split source and binary scan routing with merge-only hybrid support, binary provenance, and router-backed MCP scan contexts.

Add async MCP scan jobs for long-running scans, detector precision fixes, benchmark scripts, docs, and integration coverage.
2026-05-19 00:24:12 -04:00
trvon
3f1234f7bb
fold: bounded symbolic SAST for C++
Clang LibTooling + Z3 BV-64 encoding.
11 detectors: clone, equiv, class-dup, dead, null-deref,
div-by-zero, int-overflow, oob, int-truncation,
sign-conversion, taint-track.

Interfaces: CLI, MCP (JSON-RPC 2.0), LSP.
Output: SARIF 2.1.0, JSON, Markdown.
2026-05-14 23:14:54 -04:00