.ie \n(.g .ds Aq \(aq
.el .ds Aq '
.TH agent 1 "agent "
.SH NAME
agent \- Agent\-oriented commands (schema\-only output for AI agent prompts)
.SH SYNOPSIS
\fBmurk agent\fR [\fB\-h\fR|\fB\-\-help\fR] <\fIsubcommands\fR>
.SH DESCRIPTION
Agent\-oriented commands (schema\-only output for AI agent prompts)
.SH OPTIONS
.TP
\fB\-h\fR, \fB\-\-help\fR
Print help
.SH SUBCOMMANDS
.TP
agent\-plan(1)
Emit schema\-only context safe to paste into an AI agent prompt
.TP
agent\-exec(1)
Run a command with strict agent\-safe defaults (clears the inherited environment, strips MURK_KEY, requires \-\-only)
.TP
agent\-grant(1)
Mint a short\-lived ephemeral key that can read only the named secrets
.TP
agent\-init(1)
One\-shot onboarding: optionally set the agent allow\-list, mint a scoped grant, and print how to run the agent safely
.TP
agent\-ls(1)
List active agent grants and their TTLs
.TP
agent\-revoke(1)
Revoke an agent grant and rotate the keys it could read
.TP
agent\-connect(1)
Wire `murk mcp` into an AI editor\*(Aqs MCP config, minting a scoped grant. No CLIENT auto\-detects installed editors; give one (claude, cursor, vscode, zed, gemini, omp, codex) to target it
.TP
agent\-disconnect(1)
Remove murk\*(Aqs entry from an AI editor\*(Aqs MCP config. No CLIENT clears every configured editor; `\-\-rotate` also revokes the grant and rotates its keys
.TP
agent\-help(1)
Print this message or the help of the given subcommand(s)