CLI Commands Reference¶

Complete reference for all REPL commands in the Animus interactive shell.

Overview¶

The Animus REPL accepts two kinds of input:

Natural language — Just type what you want. Animus will use its cognitive layer to understand and act.
Slash commands — Structured commands for precise control.

Important note for local model users: When using Ollama (local models), natural language mode does not execute tools automatically. Use /tool <name> for direct tool invocation. See Working with Local Models for details.

Basic Commands¶

Command	Description
`help` or `/help`	Show the help summary
`exit`	Quit Animus
`/status`	Show system status (provider, backend, memory count)
`/stats`	Show detailed memory statistics + task success rate
`/model`	Show current model routing (primary + fallback)
`/auto`	Toggle auto-approve for tool execution

Examples¶

>>> /status
# Shows data directory, model provider, memory backend, total memories

>>> /model
# Primary: anthropic/claude-sonnet-4-20250514
# Local:   ollama/llama3:8b
# Dual-model routing active.

>>> /auto
# Toggles ANIMUS_AUTO_APPROVE environment variable

Memory Commands¶

Command	Description
`/remember <text>`	Store a semantic memory
`/recall <query>`	Search memories by semantic similarity
`/forget <id>`	Delete a memory by ID (or partial ID prefix)
`/review <id>`	Show full details of a single memory
`/list [type]`	List recent memories. Optional type: `semantic`, `episodic`, `procedural`, `active`
`/history`	Show recent conversation sessions

Examples¶

>>> /remember User prefers dark mode interfaces
# Remembered: a1b2c3d4

>>> /recall dark mode
# a1b2c3d4 [semantic] User prefers dark mode interfaces

>>> /review a1b2c3d4
# Full memory details: type, tags, confidence, created date, content, metadata

>>> /list semantic
# Lists semantic memories sorted by most recently updated

Tagging Commands¶

Command	Description
`/tag <id> <tag1> [tag2...]`	Add tags to a memory
`/untag <id> <tag>`	Remove a tag from a memory
`/tags`	List all tags with usage counts
`/search-tags <tag1> [tag2...]`	Find memories that have ALL specified tags

Examples¶

>>> /tag a1b2c3d4 preference ui
# Tagged a1b2c3d4 with: preference, ui

>>> /search-tags preference ui
# Shows memories tagged with both "preference" AND "ui"

Structured Memory¶

Command	Description
`/fact <subject> \\| <predicate> \\| <object>`	Store a structured semantic fact
`/procedure <name> \\| <trigger> \\| <step1>; <step2>...`	Store a learned workflow

Examples¶

>>> /fact User | prefers | dark mode
# Fact stored: e5f6g7h8

>>> /procedure morning-routine | waking up | check email; review calendar; plan day
# Procedure stored: i9j0k1l2

Export / Import / Backup¶

Command	Description
`/export [path]`	Export all memories to JSON (default: `~/animus-export.json`)
`/import <path>`	Import memories from a JSON file
`/backup [path]`	Create a full ZIP backup (default: `~/animus-backup.zip`)

Examples¶

>>> /export /tmp/animus-export.json
# Exported to: /tmp/animus-export.json

>>> /import ~/animus-export.json
# Imported 42 memories

Tools¶

Command	Description
`/tools`	List all available tools with parameters
`/tool <name> [param=value ...]`	Execute a tool directly

Direct Tool Execution¶

When the agent loop is disabled (Ollama provider), /tool is the only way to execute tools.

>>> /tool get_datetime
# Tool output: 2026-07-01 14:30:00

>>> /tool read_file path=/etc/hostname
# Tool output: myhostname

>>> /tool web_search query="Python context managers"
# Tool output: [search results]

Tool names with spaces: Use positional arguments for simple tools:

>>> /tool web_search Python context managers
# params["query"] = "Python context managers"

Available built-in tools: See the Tools Reference for the full list.

Reasoning Modes¶

Command	Description
`/deep <query>`	Use deep reasoning mode (step-by-step, multi-perspective)
`/research <query>`	Research mode with automatic web search
`/brief [topic]`	Generate a situation briefing from memory context

Examples¶

>>> /deep Should we migrate from SQLite to PostgreSQL?
# Animus thinks step-by-step before responding

>>> /research vector databases for semantic search
# Runs web search, then synthesizes findings

>>> /brief
# Generates morning briefing from recent memories and nudges

Decision Support¶

Command	Description
`/decide <question>`	Run structured decision analysis with pros/cons/scoring

Example¶

>>> /decide Should we use FastAPI or Flask for the API layer?
# Outputs structured analysis with dimensions, options, and recommendation

Task Tracking¶

Command	Description
`/task add <description>`	Add a new task
`/task list [--all]`	List pending tasks (add `--all` for completed too)
`/task done <id>`	Mark a task complete
`/task start <id>`	Mark a task in progress
`/task delete <id>`	Delete a task

Examples¶

>>> /task add Review pull request #105
# Task added: t1a2b3c4 - Review pull request #105

>>> /task list
# [pending]   t1a2b3c4 Review pull request #105

>>> /task done t1a2b3c4
# Task completed: t1a2b3c4

API Server (Phase 3)¶

Command	Description
`/server start [port]`	Start the FastAPI server (default: 8420)
`/server stop`	Stop the API server
`/server status`	Show server status

Voice (Phase 3)¶

Command	Description
`/voice on`	Enable voice input (microphone listening)
`/voice off`	Disable voice input
`/speak <text>`	Speak text aloud using TTS
`/speak-toggle`	Toggle automatic TTS for all responses

Dependencies: Install with pip install 'animus[voice]'

Integrations (Phase 4)¶

Command	Description
`/integrations`	List all integrations with connection status
`/integrate <service>`	Connect an integration (interactive prompts for credentials)
`/disconnect <service>`	Disconnect an integration

Available services: filesystem, todoist, google_calendar, gmail, webhooks

Example¶

>>> /integrate todoist
# Prompts for API key, then connects

Self-Learning (Phase 5)¶

Command	Description
`/learning`	Show learning dashboard (total learned, pending approval, guardrails)
`/learning scan`	Trigger pattern detection across memory
`/learning approve <id>`	Approve a pending learning
`/learning reject <id>`	Reject a pending learning
`/learning history`	Show recent learning events
`/learning rollback`	List rollback checkpoints
`/learning rollback <id>`	Rollback to a checkpoint (unlearns items)
`/unlearn <id>`	Remove a specific learned item
`/guardrails`	List all active guardrails
`/guardrail add <rule>`	Add a user-defined guardrail

Examples¶

>>> /learning
# Total learned: 12
# Pending approval: 3
# Guardrail blocks: 0

>>> /learning scan
# Detected 2 patterns
#   [habit] User reviews docs before commits (confidence: 85%)

>>> /guardrail add "Never commit API keys to git"
# Added guardrail: g-abc123def456

Citizen Zero¶

Command	Description
`/reflect`	Generate reflection candidates from session (requires approval to persist)
`/eval`	Generate eval report with owner scoring

Requires: citizen_zero.enabled = true in config.

Proactive Intelligence¶

Command	Description
`/briefing`	Generate morning briefing from nudges and memory
`/nudges`	Show active nudges with priorities
`/nudges dismiss [id]`	Dismiss one nudge (or all if no ID provided)
`/meeting-prep <topic>`	Prepare context for a meeting

Example¶

>>> /meeting-prep quarterly review
# Generates panel with relevant memories, tasks, and entities

Entities & Relationships¶

Command	Description
`/entities`	List tracked entities (people, projects, organizations, etc.)
`/entity add <name> <type> [alias1,alias2]`	Add an entity
`/entity <name>`	Show entity details and related context
`/entity delete <name>`	Delete an entity

Valid types: person, project, organization, place, topic, event, tool

Example¶

>>> /entity add Animus project
# Entity added: Animus (project)

>>> /entity Animus
# Shows mention count, last mentioned date, aliases, related memories

Cross-Device Sync (Phase 6)¶

Command	Description
`/sync start`	Start sync server and device discovery
`/sync stop`	Stop sync server and discovery
`/sync status`	Show sync status and connected peers
`/sync discover`	List discovered devices on the local network
`/sync connect <addr>`	Connect to a peer (ws://host:port or device_id)
`/sync disconnect`	Disconnect from current peer
`/sync now`	Trigger manual sync with connected peer
`/sync pair`	Show pairing code (shared secret)

Dependencies: Install with pip install websockets zeroconf

Forge (Multi-Agent Workflows)¶

Command	Description
`/forge run <path>`	Run a workflow YAML file
`/forge resume <name>`	Resume a paused/failed workflow
`/forge status [name]`	Show workflow status (or list all)
`/forge list`	List all workflows with checkpoints
`/forge pause <name>`	Pause a running workflow

Example¶

>>> /forge run workflows/examples/build_task.yaml
# Running workflow: build_task
# Agents: 3 | Gates: 2 | Budget: $2.00
# [green]Workflow complete: completed[/green]

Agent Mode¶

Command	Description
`/build <description>`	Autonomous build pipeline (plan → code → lint → test → fix)

Example¶

>>> /build add a health check endpoint
# Build Pipeline: add a health check endpoint
# Steps: planner → coder → verifier → fixer
# Budget: $2.00

Command Quick Reference Table¶

Category	Commands
Basic	`exit`, `help`, `/status`, `/stats`, `/model`, `/auto`
Memory	`/remember`, `/recall`, `/forget`, `/review`, `/list`, `/history`
Tags	`/tag`, `/untag`, `/tags`, `/search-tags`
Structured	`/fact`, `/procedure`
Export	`/export`, `/import`, `/backup`
Tools	`/tools`, `/tool`
Reasoning	`/deep`, `/research`, `/brief`
Decision	`/decide`
Tasks	`/task add/list/done/start/delete`
Server	`/server start/stop/status`
Voice	`/voice on/off`, `/speak`, `/speak-toggle`
Integrations	`/integrations`, `/integrate`, `/disconnect`
Learning	`/learning`, `/learning scan/approve/reject/history/rollback`, `/unlearn`, `/guardrails`, `/guardrail add`
Citizen Zero	`/reflect`, `/eval`
Proactive	`/briefing`, `/nudges`, `/nudges dismiss`, `/meeting-prep`
Entities	`/entities`, `/entity add/delete/<name>`
Sync	`/sync start/stop/status/discover/connect/disconnect/now/pair`
Forge	`/forge run/resume/status/list/pause`
Build	`/build`

CLI Commands Reference¶

Overview¶

Basic Commands¶

Examples¶

Memory Commands¶

Examples¶

Tagging Commands¶

Examples¶

Structured Memory¶

Examples¶

Export / Import / Backup¶

Examples¶

Tools¶

Direct Tool Execution¶

Reasoning Modes¶

Examples¶

Decision Support¶

Example¶

Task Tracking¶

Examples¶

API Server (Phase 3)¶

Voice (Phase 3)¶

Integrations (Phase 4)¶

Example¶

Self-Learning (Phase 5)¶

Examples¶

Citizen Zero¶

Proactive Intelligence¶

Example¶

Entities & Relationships¶

Example¶

Cross-Device Sync (Phase 6)¶

Forge (Multi-Agent Workflows)¶

Example¶

Agent Mode¶

Example¶

Command Quick Reference Table¶

See Also¶