ai-maestro

Share files between agents, hosts and people through an AFP space (S3-compatible storage such as Garage). Store a file and get its reference, fetch a file someone else stored, list files, make a time-limited download link. Use when the user asks to share, send, upload, publish or download a file across agents or hosts, when a message carries an afp:// reference or a reference object, and for files too large or too long-lived for a message attachment.

Authenticate an AI agent with an auth server using the Agent Identity (AID) protocol, with Ed25519 identity documents, proof of possession, OAuth 2.0 token exchange and scoped JWTs. Use when an agent must sign in to an AID-enabled server, get or refresh an access token, or prove its identity to an API. Self-contained; works without other protocols.

Send, read and reply to signed messages between AI agents with the Agent Messaging Protocol (AMP). Use when the user asks to message, notify, ask or reply to another agent, check the inbox, or read a message, and when a [MESSAGE] notification arrives. Supports local and federated delivery, file attachments and Ed25519 signatures; works with any agent that can run shell commands.

Creates, manages, and orchestrates AI agents using the AI Maestro CLI. Use when the user asks to "create agent", "list agents", "delete agent", "rename agent", "hibernate agent", "wake agent", "install plugin", "show agent", "export agent", "restart agent", "install marketplace", or any agent lifecycle management task.

Create and update interactive HTML pages (canvases) that the user sees in the AI Maestro dashboard, and handle what they click, submit or select. Use when the user asks to see something visually (a dashboard, chart, table or report), asks for a form, picker, approval or review flow, or when a [CANVAS] interaction notification arrives. Not for one-line answers, code or terminal commands.

Look up a function, class or API in this codebase's generated documentation: its signature, arguments, parameters, return type and usage. Use when asked what arguments or parameters something takes or what it returns, before calling or extending code you have not read in this session, and when the user asks to "search docs", "find function" or "check API".

Query the code graph for callers, dependencies and the blast radius of a change. Use before changing the signature or behaviour of a function, class or module that other code uses, when the user asks "find callers", "what uses this" or "check dependencies", and when exploring an unfamiliar codebase.

Search this agent's long-term memory of past sessions: decisions, facts, corrections, and what depends on an entity. Use when a task touches something with history (a repo, service, environment or decision from earlier sessions), before deploying, moving, deleting or reconfiguring something to see what depends on it (memory-search.sh --about <entity>), and when the user asks "what did we discuss", "remember when", "search memory", "find previous conversation" or "check history".

Creates and manages persistent markdown planning files (task_plan.md, findings.md, progress.md) for complex task execution. Use when starting multi-step tasks, research projects, or any task requiring >5 tool calls. Solves the EXECUTION problem - staying focused during long-running tasks.