Cette page est affichée en anglais. Une traduction française est en cours.
CLI ToolsMay 14, 2026·2 min de lecture
Augustus — LLM Vulnerability Scanner (Go CLI)
Augustus is a Go-based LLM vulnerability scanner covering 210+ adversarial attacks and 28 providers; verified 205★ and pushed 2026-05-11.
Prêt pour agents
Staging sûr pour cet actif
Cet actif est d'abord staged. Le prompt copié demande à l'agent d'inspecter les fichiers staged avant d'activer scripts, config MCP ou config globale.
Stage only · 17/100Policy : staging
Surface agent
Tout agent MCP/CLI
Confiance
Confiance : Established
Commande de staging sûr
npx -y tokrepo@latest install 23d77068-dbd1-55d2-a53d-fc6a5f3929d7 --target codexStage les fichiers d'abord; l'activation exige la revue du README et du plan staged.
Introduction
Augustus is a Go-based LLM vulnerability scanner covering 210+ adversarial attacks and 28 providers; verified 205★ and pushed 2026-05-11.
Best for: Security teams running repeatable prompt-injection/jailbreak regression tests across multiple LLM providers
Works with: Go 1.25.3+ and credentials for whichever LLM provider you target
Setup time: 6-15 minutes
Key facts (verified)
- GitHub: 205 stars · 25 forks · pushed 2026-05-11.
- License: Apache-2.0 · owner avatar + repo URL verified via GitHub API.
- README-backed entrypoint:
augustus --help.
Main
Attack coverage: README claims 210+ adversarial attacks across prompt injection, jailbreak, encoding exploits, and data extraction.
Provider breadth: README states integration with 28 LLM providers, useful for comparing models and configs under one harness.
Ops-friendly scanning: README highlights concurrency, rate limiting, retries, and timeouts for production-style testing.
Reportable outputs: README includes sections on output formats so you can wire results into CI/security review.
Source-backed notes
- README describes Augustus as a Go-based LLM vulnerability scanner for security professionals.
- README quick start uses
go install .../cmd/augustus@latest and requires Go 1.25.3+.
- README warns some probes use offensive language for jailbreak testing; use only in authorized environments.
FAQ
- Is it safe to run against production?: Treat it like a security test tool: rate-limit, scope targets, and use authorized environments only.
- Do I need provider API keys?: Yes — you need credentials for each LLM provider you want to scan.
- How do I start small?: Run a single probe first, then expand to multiple probes and report outputs once you trust the harness.
Fil de discussion
Connectez-vous pour rejoindre la discussion.
Aucun commentaire pour l'instant. Soyez le premier à partager votre avis.