Superagent

Agent Vault

Infisical's open-source credential broker that proxies API calls for AI agents so they never hold secrets themselves.

ThinkWatch Lite

A local gateway that sits between your coding agents and model providers, letting you switch upstreams, redact secrets before they leave your machine, block dangerous tool calls and see the cost of each request.

OpenAPPA

OpenAPPA is a deterministic guardrail that sits between an AI agent and its tools and tracks how data flows across tool calls, instead of asking a second model to judge each call. Policies decide whether untrusted or sensitive data may reach a given action, which blocks data exfiltration caused by prompt injection or hallucination while letting agents finish their tasks. It plugs into Claude Code, Pi, OpenCode, kAgent and Archestra, comes with reusable policy "batteries", validation, observability and reporting, and is backed by a NeurIPS workshop paper with published benchmarks. It is in preview as an open RFC.

Wren Pre-sign Check

Wren Pre-sign Check is a pay-per-call x402 API that an AI agent calls before it signs anything with its wallet. Send a transaction (to + data), an EIP-712 typed-data payload or untrusted text, and it returns block, review or allow, the decoded call in plain words and every finding: unlimited approvals, Permit and Permit2 signatures, setApprovalForAll, EIP-7702 delegation, contract deployment and injected instructions. With an optional wren-guard policy (allowlist, spenders, caps) it also blocks unknown recipients, over-cap amounts and lookalike addresses. It is a static check and stores nothing you send; free demos show real output.

Eleven

An agent-first reverse-engineering app for macOS, Windows and Linux: load any app or binary and ask your AI agent what it does while it disassembles and analyzes the code for you.

wren-guard Pro

wren-guard Pro is a TypeScript guard library for AI agents that hold or move funds on Base and other EVM chains (viem, CDP AgentKit, Privy, Eliza and similar stacks). It enforces allowlisted, capped and simulated sends, decodes calldata and signatures to flag unlimited approvals, Permit, Permit2, setApprovalForAll and EIP-7702 delegation, screens for injected instructions and watches nonces. It ships with an 89-scenario attack corpus drawn from public agent wallet incidents and a GitHub Action that fails CI if your agent's policy would let a known attack through. The original Python wren-guard is free and MIT licensed.