메뉴
Run, inspect, and configure AI agents in a Memory project — trigger runs, monitor status, list runs, respond to questions, manage agent definitions and schedules.
Codex 또는 Claude로 설치 이 Prompt를 복사해 Codex, Claude 또는 다른 어시스턴트에 붙여 넣으면 Skill 페이지를 검토하고 설치를 진행할 수 있습니다.
Full Memory CLI command reference with all subcommands and flags. Use when you need exact command syntax, flag names, or usage examples for any `memory` CLI command.
Full Memory CLI command reference with all subcommands and flags. Use when you need exact command syntax, flag names, or usage examples for any `memory` CLI command.
Create a versioned release — commit, tag, push, and deploy to servers. Use when the user wants to release, deploy, publish, or push a new version.
Analyze and dump OpenCode sessions from any project. Use when the user asks to read a session, find a session by title, list sessions for a project, dump session content, or show usage stats (skills, tools, tokens, cost) for any project directory.
Read from the knowledge graph — natural language questions, semantic search, or hybrid search over objects and documents.
Create and manage graph branches — isolated workspaces for staging changes before merging to main. Use for what-if analysis, planning, or any work that should not immediately affect the main graph.
SOC 직업 분류 기준
| name | memory-agents |
| description | Run, inspect, and configure AI agents in a Memory project — trigger runs, monitor status, list runs, respond to questions, manage agent definitions and schedules. |
| metadata | {"author":"emergent","version":"2.0"} |
Manage runtime agents and agent definitions in an Emergent project using memory agents and memory agent-definitions.
.env.local containing MEMORY_PROJECT or MEMORY_PROJECT_ID. If .env.local is present anywhere above the current directory, --project is not needed. Only pass --project <id> explicitly when overriding or when no .env.local exists.agent-definitions) — the blueprint: system prompt, model config, tools, flow type, visibility. Created once, reused.agents) — a live instance of a definition in a project, with scheduling, triggers, and run history.memory agent-definitions list
memory agent-defs list --output json
memory agent-definitions get <definition-id>
memory agent-definitions create --name "My Agent" --description "Does X"
memory agent-definitions update <definition-id> --name "New Name"
memory agent-definitions delete <definition-id>
memory agents list
memory agents list --project <id> --output json
memory agents get <agent-id>
memory agents create --name "My Agent" --definition-id <def-id>
memory agents trigger <agent-id>
memory agents runs <agent-id>
memory agents runs <agent-id> --limit 20
memory agents update <agent-id> --name "New Name"
memory agents delete <agent-id>
Agents can pause and ask clarifying questions. Use these commands to list and respond.
memory agents questions list <run-id>
memory agents questions list-project
memory agents questions list-project --project <id>
memory agents questions respond <question-id> --answer "Yes, proceed"
memory agents hooks list <agent-id>
trigger kicks off an immediate run; check runs to see statusruns --limit 50 --output json for scripted monitoringquestions list-project to find pending questions, then questions respond to unblock itlist --output json and look up by nameagent-definitions has aliases: agent-defs, defs--project global flag selects the project; falls back to config default