-
Notifications
You must be signed in to change notification settings - Fork 0
Home
hyperpolymath edited this page Jun 1, 2026
·
3 revisions
panic-attack is a Rust CLI providing static analysis, taint tracking, cross-language vulnerability detection, and CVE triage across 49 languages in 500+ repositories. It operates in three deployment modes (standalone binary, panicbot CI integration, mass-panic batch scanning) and exposes 25 weak-point categories with a miniKanren-based logic engine.
The Wiki is the orientation surface; canonical docs live in the repository. Use the sections below to jump to what you need.
| If you want to… | Start here |
|---|---|
| Install + first scan | Installation · Quickstart (user) |
| Understand what it detects | Detection Categories · Languages |
| Integrate with CI | panicbot setup · pre-commit hook |
| Plug new findings into VeriSimDB | VeriSimDB Persistence |
| Triage CVE phantoms | Patch Bridge |
| Run estate-scale analysis | mass-panic · Chapel Metalayer |
| Develop / contribute | Quickstart (dev) · Architecture · CONTRIBUTING.md |
- README.adoc — top-level overview
- ROADMAP.adoc — current + planned milestones
- CHANGELOG.md — release history
- DESIGN.md — architecture rationale
- VISION.md — long-horizon direction
- EXPLAINME.adoc — verbose explanation surface
- SECURITY.md — vulnerability reporting
- PROOF-NEEDS.md · TEST-NEEDS.md — verification debt
-
0-AI-MANIFEST.a2ml— top-level AI/agent manifest (canonical-locations, invariants, capability tags) -
.machine_readable/6a2/{STATE,META,ECOSYSTEM}.a2ml— RSR-canonical clade docs -
.machine_readable/CLADE.a2ml·.machine_readable/anchors/ANCHOR.a2ml— gv-clade-index identity -
.machine_readable/agent_instructions/{coverage,debt,methodology}.a2ml— per-aspect agent guidance -
.machine_readable/integrations/{verisimdb,proven,vexometer,feedback-o-tron}.a2ml— partner contracts
- Version: 2.5.0 (CRG C / Beta)
- Tests: 282+ passing across unit / property / e2e / aspect / integration tiers
-
Languages: 49 (
.thyIsabelle +.vCoq added in v2.1.0) - Categories: 25 weak-point types (PA001–PA025 spanning resource exhaustion, crypto misuse, proof drift, supply-chain, input-boundary, mutation-gap)
- Last cleanup: 2026-06-01 — baseline-red corrective maintenance landed via #94/#97; rsr-template gap fills via #96; dependabot bumps via #93
-
Chapel Wave 2:
chapel-multilocale(7th strict gate) landed via #99 —mass-panic --numLocales=2overgasnet+smpsingle-host oversubscribed, source-built + cached. See Chapel-Metalayer for the full toolchain story.
- v2.2.0 — VeriSimDB Integration: hexad persistence for Patch Bridge mitigation registry, historical trend queries via VCL
- v2.4.0 — Patch Bridge Phase 2: hexad lifecycle persistence, auto-retire, upstream health monitoring, multi-lockfile support (beyond Cargo.lock)
- v2.5.x — Detection coverage: long-tail categories — test-context awareness, JIT context, FFI subtyping, attack-surface widening completions
See ROADMAP.adoc for the full list.
panic-attack is part of the hyperpolymath ecosystem:
-
gitbot-fleet consumes findings via
panicbottranslator (PA021/PA022/PA023/PA024/PA025 wired) -
hypatia consumes JSON
AssailReportvia Elixir rules (Logtalk export removed 2026-04-12) -
VeriSimDB persists scan octads + hexads via REST (ureq v3 /
verisim-panic-apion Fly.io) - echo-types / kategoria / typed-wasm — adjacent verification stack
-
standards — reusable CI workflows (
rust-ci-reusable.yml+ governance)
- Bugs: GitHub issues
-
Security: see SECURITY.md —
j.d.a.jewell@open.ac.uk - Discussions: GitHub discussions
panic-attack
Detection
Integration
Scale
Development