Aget Agent

From GM-RKB
(Redirected from AGET Agent)
Jump to navigation Jump to search

An Aget Agent is a Configured Conversational AI Agent that is instantiated from an AGET Framework template and keeps a persistent, governed identity (memory, skills, lessons) across sessions and AI systems, on behalf of a human principal.

  • AKA: Aget, Aget Instance.
  • Context:
    • It can be told apart from "AGET" in capitals, which names the AGET Framework rather than an agent built from it.
    • It can (typically) conform to a specific AGET Framework release (such as v3.36.0), recorded in its version file, and migrate when a new release ships.
    • It can (typically) stop at explicit decision points for its principal's approval, because AGET is not an autonomous system.
    • It can keep its knowledge in a repository that its principal holds, as a shared knowledge base rather than hidden AI state, so that the same Aget can be run through different CLI agents, such as Claude Code, Codex CLI or Gemini CLI.
    • It can be one of a fleet of Agets coordinated by a Supervisor AGET, with clear authority boundaries between them.
    • It can ask another Aget for work through an Aget Work Request.
    • It can be a Worker AGET or a Supervisor AGET, depending on its role in its fleet.
    • ...
  • Example(s):
    • Worker AGETs, which execute domain-specific tasks within a fleet, such as:
      • a content-enhancement Aget that maintains pages in a wiki.
    • Supervisor AGETs, which govern a fleet of worker Agets.
    • Agets built from the framework's role templates, such as an advisor Aget built from the advisor template for advisory work.
    • ...
  • Counter-Example(s):
    • The AGET Framework itself, which is the framework that Agets are built from rather than an agent.
    • An OpenAI CustomGPT-based Chatbot, a sibling configured conversational agent whose configuration lives in one vendor's workspace, so it does not carry an identity across AI systems.
    • An AI coding agent run without an AGET template, which starts each session without an Aget's memory, skills and lessons.
    • ...
  • See: AGET Framework, Configured Conversational AI Agent, Personal AGI System, Aget Work Request.

References

2026

  • AGET Framework. (2026). aget-framework/aget README. Retrieved 2026-10-15; no archive snapshot confirmed. Supports persistence, supervision, platforms, fleets and templates: "It gives your agents persistent knowledge, shared memory, fleet coordination, and human-supervised governance."; "Shared Memory: KB as collaboration substrate, not hidden AI state"; "Open in your CLI agent (Claude Code, Codex CLI, or Gemini CLI)"; "Configure identity in .aget/version.json"; "Gate Discipline: Explicit decision points with human approval"; "Fleet Coordination: Multi-agent patterns with clear authority boundaries"; "Not an autonomous system. AGET requires human supervision and gate discipline."; template-advisor-aget, "Advisory with personas"; "Your CLI agents lose context between sessions."
  • AGET Framework. (2026). AGET Framework releases. Retrieved 2026-10-15; no archive snapshot confirmed. Supports the release example and migration: "v3.36.0 — Fleet migration kit for supervisors", released 2026-10-10 and marked Latest.