Security●●●Banger
Raypher–eBPF-based runtime security and hardware identity for AI agents
eBPF kernel hooks enforce agent boundaries at <0.05ms latency; no API polling tax.
Big BrainBold BetWizardry
Kidiga
213mo ago
Back to browse
GitHub Repository
A secure* runtime for autonomous AI agents. Policy from plain-English constitutions. (*https://ironcurtain.dev)
498 starsTypeScript
IronCurtain: A secure* runtime for AI agent loops
by nielsprovos·Mar 3, 2026·1 point·1 comment
AI Analysis
●●●BangerBig BrainZero to OneSolve My Problem
Sandbox agents via natural-language policy, not ambient authority—genuinely novel approach.
Strengths
- •Compiles English constitutions into deterministic, enforceable security policies without LLM drift assumptions.
- •Dual-mode (auto-approve + manual escalation) balances autonomy and safety—solves the sandbox vs. trust tradeoff.
- •Active validation against test scenarios; security doesn't depend on model behavior.
Weaknesses
- •Research prototype; APIs and architecture unstable, may change significantly.
- •Threat model assumes LLM compromise but unclear how policy protects against sophisticated prompt injection within the constitutions themselves.
Category
Target Audience
AI/ML engineers, autonomous agent builders, anyone deploying untrusted LLM-based systems
Similar To
Anthropic's Constitutional AI · LangChain's tool-calling guards · Pydantic validators for LLM outputs
Similar Projects
Security●●●Banger
Telos – eBPF/LSM Runtime Security for Autonomous AI Agents
Kernel-level intent tracking stops AI exfiltration where EDR and Docker fail.
WizardryBig BrainBold Bet
nevinshine
103mo ago
AI/ML●●●Banger
Arden – Runtime policy enforcement and governance for AI agents
Intercepts tool calls before execution to block dangerous actions like DB deletes.
Solve My ProblemBold Bet
rishabtandon
851mo ago
Security●●Solid
Burrow – Runtime Security for AI Agents
Natural language policies block risky agent actions before they execute.
Solve My ProblemSlick
saranshrana
302mo ago
Security●●●Banger
Crawdad – Runtime security layer for autonomous AI agents
Post-quantum crypto and ZK proofs for AI agent security when the category is still emerging.
Big BrainSolve My Problem
AndrewGS
102mo ago
Security●●Solid
Runtime security for AI agents(injection,tool abuse, data exfiltration)
OPA-based policy engine for AI agents blocking injection and tool abuse.
Solve My Problem
dshapi
121mo ago