Detector Reference
Chainvet's static analysis ships 48 detectors across 8 categories, each with a stable ID, a default confidence, and a severity.
Chainvet's static-analysis engine ships 48 detectors organized into 8
categories. Each detector has a stable ID (e.g. AC-01, RE-04), a
lowercase slug (e.g. arbitrary-transfer-from), a severity (impact), and a
default confidence (precision). The taxonomy — 47 of these detectors — is the
canonical rule set advertised as "45+ detectors".
Categories
| Prefix | Category | Detectors | Page |
|---|---|---|---|
AC | Access Control | 18 | Access Control |
AR | Arithmetic | 4 | Arithmetic |
BM | Block Manipulation | 3 | Block Manipulation |
CR | Cryptographic | 2 | Cryptographic |
DS | Denial of Service | 7 | Denial of Service |
RE | Reentrancy | 5 | Reentrancy |
SM | Storage & Memory | 7 | Storage & Memory |
MI | Miscellaneous | 2 | Miscellaneous |
How to read a detector
Every finding a detector produces carries two ratings — read them together:
- Severity — the potential impact (
high/medium/low), fixed by the detector per finding. - Confidence — the detector's own precision estimate (
high/medium/low). A detector raises or lowers it per finding when it has local evidence; otherwise it uses a per-kind default. The values in these tables are that default.
See Severity & Confidence for the full model,
and note that some detectors emit different severities depending on context (for
example, arithmetic overflow is medium in an unchecked block but high on
untrusted input in Solidity < 0.8) — such cases are shown as a range.
These run in every mode
The detectors are the static-analysis layer. They run on their own in -m static
and also participate in hybrid mode, where symbolic execution and fuzzing confirm
and extend their findings. See Analysis engines.