Menu
Store, search, and manage shared knowledge through the memory service. Use for persisting information across agent sessions, semantic search over stored knowledge, and inter-agent knowledge sharing.
Install with Codex or Claude Copy this prompt, paste it into Codex, Claude, or another assistant, and let it review the skill page and install it for you.
Based on SOC occupation classification
Browse, read, and author per-project knowledgebase documents through the knowledge service CLI. Use for creating, updating, listing, and reading markdown documents stored against a project, and reconciling DB rows against disk.
Cut a new agentd release. Suggests the next semver version from conventional commits (or accepts an explicit `x.y.z`), bumps the workspace version, regenerates the changelog with git-cliff, then commits, tags, and pushes the `v*` tag that triggers the Release workflow. Use when the user runs /release or asks to publish/cut/tag a new version.
Create a new GitHub plan issue with the "Plan: <title>" format, appropriate labels, and structured body. Use when the user wants to capture a high-level plan for future work.
Ask the human user a question during workflow execution and check for answers. Use for agent-to-human Q&A, collecting decisions, and reacting to human responses via ask_response workflow triggers.
Branch stacking and PR management with git-spice for the agentd project. Use when creating stacked branches, submitting PRs, syncing with upstream, resolving rebase conflicts, or navigating a branch stack.
Create, list, and manage conversation rooms, participants, and messages through the communicate service. Use for inter-agent messaging, room management, real-time message watching, and participant coordination.
| name | agent-memory |
| description | Store, search, and manage shared knowledge through the memory service. Use for persisting information across agent sessions, semantic search over stored knowledge, and inter-agent knowledge sharing. |
Skill for interacting with the agentd memory service — a vector-backed knowledge store for agents.
# Store a simple fact
agent memory remember "The deployment key is stored in 1Password vault 'Infrastructure'"
# With metadata
agent memory remember "OAuth tokens expire after 24 hours" \
--created-by worker-agent \
--type information \
--tags auth,tokens,expiry
# Store a question for later resolution
agent memory remember "Should we migrate to async Redis client?" \
--type question \
--tags redis,architecture
# Store a request
agent memory remember "Need someone to review the auth middleware rewrite" \
--type request \
--tags review,auth
# Control visibility
agent memory remember "API key rotation schedule: first Monday of each month" \
--visibility shared \
--share-with ops-agent,deploy-agent
# Private memory (only visible to creator)
agent memory remember "My working hypothesis: the leak is in the connection pool" \
--visibility private \
--created-by debug-agent
# Reference other memories
agent memory remember "Follow-up: the Redis migration decision was made" \
--references <memory-id-1>,<memory-id-2>
information — facts, decisions, observations (default)question — open questions needing resolutionrequest — action items or askspublic — visible to all agents (default)shared — visible to creator and specified actorsprivate — visible only to creatorSemantic similarity search over all stored memories:
# Basic search
agent memory search "how do we handle token expiry"
# Search as a specific actor (respects visibility)
agent memory search "deployment procedures" --as-actor deploy-agent
# Filter by type and tags
agent memory search "auth" --type information --tags auth
# Limit results
agent memory search "Redis" --limit 5
# Time-bounded search
agent memory search "incidents" --since 2026-01-01T00:00:00Z --until 2026-03-01T00:00:00Z
# JSON output
agent memory search "auth" --json
# Get a specific memory by ID
agent memory recall <memory-id>
# List memories with filters
agent memory list
agent memory list --type question
agent memory list --tag auth
agent memory list --created-by worker-agent
agent memory list --visibility public
agent memory list --limit 20 --offset 40
# Make a memory shared with specific agents
agent memory visibility <memory-id> shared --share-with ops-agent,deploy-agent
# Make public
agent memory visibility <memory-id> public
# Make private
agent memory visibility <memory-id> private
agent memory forget <memory-id>
agent memory health
When running as an agent with the agentd MCP server connected, the core operations are
also available as tools (no CLI needed): remember (store a memory), search_memories,
list_memories, and get_memory. Prefer these during autonomous execution; use the CLI
for interactive/human use, visibility changes, and deletion.
remember takes content and created_by (your agent identity) plus optional
memory_type (information | question | request), tags, visibility
(public | private | shared), shared_with, and references.