circle-ir-ai 2.7.12 → 2.7.14

This diff represents the content of publicly available package versions that have been released to one of the supported registries. The information contained in this diff is provided for informational purposes only and reflects changes between package versions as they appear in their respective public registries.
Files changed (2) hide show
  1. package/CHANGELOG.md +49 -0
  2. package/package.json +2 -2
package/CHANGELOG.md CHANGED
@@ -5,6 +5,55 @@ All notable changes to this project will be documented in this file.
5
5
  The format is based on [Keep a Changelog](https://keepachangelog.com/en/1.0.0/),
6
6
  and this project adheres to [Semantic Versioning](https://semver.org/spec/v2.0.0.html).
7
7
 
8
+ ## [2.7.14] - 2026-05-21
9
+
10
+ ### Dependencies
11
+
12
+ - circle-ir upgraded to ^3.22.2 — fixes the Rust regression we
13
+ reported at circle-ir#4 yesterday. All Rust benchmarks restored
14
+ to 3.22.0 baseline while keeping 3.22.1's Go/Bash wins.
15
+
16
+ ### Net benchmark state (3.22.2 vs 3.22.0)
17
+
18
+ | Benchmark | 3.22.0 | 3.22.2 | Δ |
19
+ |---|---|---|---|
20
+ | Go Synthetic | 54.2% | **64.2%** | +10pp |
21
+ | Vulnerability-goapp | 62.5% | **70.0%** | +7.5pp |
22
+ | Rust Synthetic | 92.3% | 92.3% | = (restored) |
23
+ | CWE-Bench-Rust | F1 97.1% | F1 97.1% | = (restored) |
24
+ | Everything else | unchanged | unchanged | = |
25
+
26
+ Clean net wins, no regressions.
27
+
28
+ ## [2.7.13] - 2026-05-21
29
+
30
+ ### Dependencies
31
+
32
+ - circle-ir upgraded to ^3.22.1.
33
+
34
+ ### Net benchmark impact (3.22.0 → 3.22.1)
35
+
36
+ **Wins:**
37
+ - Go Synthetic: 54.2% → **64.2%** (+10pp, FPR 30→20) — circle-ir 3.22.1 P1 placeholder-aware SQLi filter
38
+ - Vulnerability-goapp: 62.5% → **70.0%** (+7.5pp)
39
+ - Bash: `eval` / `bash -c` / cmd-substitution patterns now detected → closes
40
+ cogniumhq/circle-ir-ai#33
41
+
42
+ **Regressions (tracked at circle-ir#4):**
43
+ - Rust Synthetic: 92.3% → **81.5%** (TPR 100→89.2, 4 cmdi cases lost)
44
+ - CWE-Bench-Rust: F1 97.1% → **90.9%**
45
+ - SecuriBench Micro: TPR 97.7% → **96.8%** (1 TP lost)
46
+
47
+ Root cause: 3.22.1's expanded `SAFE_RECEIVERS_BY_METHOD` and loosened
48
+ `receiverMightBeClass()` heuristic over-filter Rust's
49
+ `Command::new("sh").arg("-c").arg(input).output()` pattern as if it
50
+ were the safe `exec.Command(prog, args...)` no-shell form. Filed
51
+ upstream with 4 concrete repros from `benchmarks/runners/run-rust.ts`.
52
+
53
+ Net judgment: shipping the bump anyway — Go gains affect more users
54
+ than the Rust regression, and the regression is well-localized +
55
+ reportable. Will re-evaluate when 3.22.2 lands.
56
+
8
57
  ## [2.7.12] - 2026-05-21
9
58
 
10
59
  ### Changed
package/package.json CHANGED
@@ -1,6 +1,6 @@
1
1
  {
2
2
  "name": "circle-ir-ai",
3
- "version": "2.7.12",
3
+ "version": "2.7.14",
4
4
  "description": "LLM-enhanced SAST analysis built on circle-ir",
5
5
  "main": "dist/index.js",
6
6
  "module": "dist/index.js",
@@ -94,7 +94,7 @@
94
94
  "dependencies": {
95
95
  "@ax-llm/ax": "^20.0.0",
96
96
  "@mastra/core": "^1.18.0",
97
- "circle-ir": "^3.22.0",
97
+ "circle-ir": "^3.22.2",
98
98
  "minimatch": "^10.2.5",
99
99
  "p-queue": "^9.1.0"
100
100
  },