whereismyai
Planlock
Planlock is an MCP proxy that approves an AI agent plan once and enforces per-step limits on tool calls.
Planlock is an MCP proxy designed to secure AI agents that use tools. It sits between the agent and its tools to enforce a pre-approved plan, rejecting any calls that fall outside the approved scope. Users approve the plan once through an out-of-band, token-protected approver channel. The system checks allowed tools, argument limits, maximum calls, and optional version pins. If an agent attempts an unapproved action, Planlock blocks the call and halts the execution until a human overrides it.
Planlock at a glance
Key features
- Approves multi-step agent plans once
- Enforces per-step allowed tools and argument limits
- Blocks calls that deviate from the plan
- Supports optional version pins for upstream scripts
- Provides an out-of-band token-protected approver channel
- Halts execution on failed or blocked steps
- Runs security regression tests via npm test
Use it for
- Controlling tool use in AI agents
- Preventing unauthorized financial transactions
- Restricting agent access to specific script versions
- Enforcing strict step-by-step agent workflows
Alternatives to Planlock
- mcpgawk — Pins each tool to what you approved and blocks calls if a server changes.
- AgentGate — Blocks attacks in under 35ms with a local neural classification head.
- Context Data — Authorizes agent access to production data with per-column control and auditing.
- MCP Bouncer — Stops agents from exfiltrating data to destinations found in untrusted tool output.
- Hlock — Pins entire AI agent setups into a portable file to share and run anywhere.
- Plano — Offloads critical infrastructure tasks for faster agent deployment.
- Plannotator — Review AI agent plans and code diffs visually before execution and commit.
- HaltState MCP — Enforces backend action policies automatically without human approval queues.
- OpenAPPA — Tracks data flows algebraically to prevent prompt injection without breaking agents.
- Stop That Shit — Enforces task boundaries to stop unnecessary AI agent actions.
- Paveo — Enforces deterministic rules on AI agents to prevent destructive actions.
- Plango — Protects Mac computers and AI tools with local cyber intelligence.
All Planlock alternatives