A web app that intercepts MCP tool calls and injects scoped, time-limited credentials so AI agents never hold raw API keys

93% of AI agent projects paste raw API keys directly into MCP config files, creating a credential leak surface that scales with every tool added. Developers have no native way to scope, rotate, or audit which keys an agent actually used. This web app sits as a lightweight proxy between the MCP client and the upstream APIs, injecting short-lived credentials at call time and revoking them automatically, so a compromised agent config exposes nothing.

Demand Breakdown

GitHub

1,747

659

OPENAI_FORUM

Social Proof 5 sources

Infisical/agent-vault: Open-source credential proxy and vault for agents

@infisical · 2026-04-22

1,747 HN

Show HN: Zerobox - Sandbox any command with file, network, credential controls

@afshinm · 2026-03-30

234 HN

Show HN: OneCLI - Vault for AI Agents in Rust

@onecli · 2026-03-12

213 HN

Show HN: Agent Vault - Open-source credential proxy and vault for agents

@infisical · 2026-04-22

212 OA

The State of Secrets Sprawl 2026: AI-Service Leaks Surge 81% and 29M Secrets Hit Public GitHub

@GitGuardian · 2026-03-01

Gap Assessment

CompetitiveMultiple tools exist but differentiation opportunities remain

4 tools exist (Infisical Agent Vault, OneCLI, Peta, Alter) but gaps remain: No managed SaaS tier, no per-call audit log UI, no per-agent key scoping dashboard, no revocation policy rules for non-engineering users.; CLI-only, no web dashboard, no MCP-native config injection, no usage metering, no team management..

Features8 agent-ready prompts

MCP-native proxy endpoint with zero-code config injection

▶

Per-agent credential scoping and issuance

▶

Short-lived token issuance with automatic expiry

▶

Per-call audit log with tool name, agent ID, and credential used

▶

Credential rotation with no agent downtime

▶

Policy rules engine for call-level access control

▶

Team and workspace management with role-based access

▶

One-click MCP server library with pre-built integrations

▶

Competitive LandscapeFREE

Product	Does	Missing
Infisical Agent Vault	Open-source credential proxy that intercepts MCP tool calls and injects secrets at the transport layer; requires self-hosting.	No managed SaaS tier, no per-call audit log UI, no per-agent key scoping dashboard, no revocation policy rules for non-engineering users.
OneCLI	Rust-based vault where agents authenticate with a single JWT and never receive raw keys; 2,369 GitHub stars.	CLI-only, no web dashboard, no MCP-native config injection, no usage metering, no team management.
Peta	Marketed as '1Password for AI agents'; server-side encrypted vault that issues scoped, time-limited tokens per operation.	Enterprise-focused pricing and onboarding; no lightweight self-serve tier for solo developers or small teams running open MCP servers.
Alter	Zero-trust identity and access control for AI agents; wraps every tool call in authentication and real-time authorization checks.	Broad zero-trust platform, not MCP-native; complex enterprise integration; no per-MCP-server config injection out of the box.