RepoRadar
Leave a Tip
Item detail

openguardrails/agentfw

agentfw is an MIT-licensed TypeScript local proxy that sits on the wire between any AI agent (Claude Code, Codex, OpenClaw, Hermes, Claude Desktop) and the LLMs it calls, doing two things in one place: useful routing/repair and security guardrails. It masks API keys and wallet secrets before they reach the model, runs detectors on every request/response, and ships live call inspection plus capabil

Open source →Back to radar
Score8.4
Popularity78.0
Riskconditional
TierGold
Score breakdown
Usefulness9.0
Novelty8.0
Momentum8.0
Maturity8.2
Open-source/build8.4
Evidence7.2
Workflow potential9.9
Setup ease8.8

Popularity is tracked separately. Support, ads, sponsorships, and tips never affect these signals.

Why it matters

Useful for developers running Claude Code, Codex, or any MCP-speaking agent who wants a single chokepoint to see every model call, mask secrets on the way out, and route subagents to cheaper models while the planner stays on the strong one.

Who should use it

developers running Claude Code or Codex with real API keys and wallet credentialsagent platform teams that need a single observability + security chokepointteams routing agent subagents to cheaper models while the planner stays on the strong onesecurity-conscious users who want credential masking without adopting a vendor framework

Who should skip it

Skip if the source link, docs, or setup requirements do not match your workflow.

Risk explanation

agentfw is a security chokepoint — run it carefully during upgrades and pin versions, since a misconfigured proxy can drop or rewrite legitimate traffic.

Evidence links

Closest alternatives / related signals

agent-securityproxyclaude-codecodexmcpsecretsroutingfirewall