Supported agents
Every agent Pushary supports, how it connects, and what it can enforce
Every agent can send you notifications and ask you questions. The difference is whether Pushary can enforce approvals (stop an action until you say yes) or only cooperate (the agent chooses when to ask).
| Agent | How it connects | Enforced approvals | Notifications | Setup |
|---|---|---|---|---|
| Claude Code | Native hooks + MCP | Yes. Shell, writes, and edits are blocked until you approve | Yes | npx @pushary/agent-hooks setup, pick Claude Code |
| Codex | Native hooks + MCP | Yes. Shell and apply_patch edits. Needs Codex 0.122 or newer | Yes | npx @pushary/agent-hooks setup, pick Codex |
| Cursor | Plugin, or plain MCP | Yes with the plugin. Shell commands only | Yes | npx @pushary/agent-hooks setup, pick Cursor |
| Hermes | Native Python plugin | Opt-in. Set PUSHARY_GATE_TOOLS to gate specific tools | Yes, plus automatic error alerts | npx @pushary/agent-hooks setup, pick Hermes |
| Windsurf | Plain MCP | No. The agent chooses when to ask | Yes | Setup, pick Other |
| Any MCP client | Plain MCP (mcp-remote for stdio) | No | Yes | Setup, pick Other |
| No-code and HTTP (n8n, Zapier, scripts) | REST with a Bearer token | No | Yes | Connect any agent |
A few notes from the details:
- Enforced vs cooperative. Enforced means the agent is physically stopped until you approve. Cooperative means the agent decides whether to ask, so it might not. Only Claude Code, Codex, the Cursor plugin, and Hermes with gating turned on can enforce.
- What "hooks" are. Hooks let Pushary step in before an agent runs an action. Agents without hooks (plain MCP, Windsurf, no-code) get notifications and questions only.
- Cursor is shell only. Cursor only exposes a hook for shell execution, so the gate covers terminal commands, not file reads or edits. See the Cursor guide.
- Same tools everywhere. Every MCP integration exposes the same four tools:
send_notification,ask_user,wait_for_answer,cancel_question. See the tools reference.
I want to...
| Goal | Where to go |
|---|---|
| Get pinged when my agent finishes | Receiving notifications |
| Approve risky commands before they run | Permission policies |
| Answer my agent's questions from my phone | Human in the loop |
| Stop everything right now | Kill switch |
| See what my agents did | Audit log |
| Connect a tool not listed above | Connect any agent |