Skip to main content

​
help

Display all available commands and usage information. Usage: 
help
Output: 
OneClaw Commands:
  ask Q        - Ask AI a question
  tools        - List registered tools
  tool X k=v   - Execute tool X with params (key=value)
  channels     - List active channels
  events       - Show event bus status and recent events
  status       - Show agent status and config
  metrics      - Show operational metrics (counters, uptime)
  health       - Health check all layers
  reload       - Check config file for changes (report only)
  providers    - List LLM providers and status
  pair         - Generate device pairing code
  verify CODE  - Pair device with 6-digit code
  devices      - List all paired devices
  unpair ID    - Remove a paired device (prefix match)
  remember X   - Store X in memory
  recall X     - Search memory for X
  help         - Show this help message
  exit         - Shut down the agent

Any other input will be processed by the AI pipeline.
Security: Always open (no authorization required)

​
status

Display system overview with runtime configuration and current state. Usage: 
status
Example Output: 
OneClaw Agent v1.5.0

  Uptime: 2h 15m 30s

  Security: enforced
    Memory: 42 entries (sqlite)
    Embedding: nomic-embed-text (768d) — 38 embedded / 42 total
    Provider: anthropic (online)
    Chain: anthropic → ollama (fallback)
    Tools: 3
    Events: 1250 processed, 0 pending
    Messages: 87 total (2 denied)
    LLM: 15 calls (avg 245ms)

  Type 'health' for detailed check
    Type 'metrics' for full telemetry
Security: Requires system:status authorization Fields:
  • Uptime: Time since runtime started
  • Security: enforced (deny-by-default) or open
  • Memory: Entry count and backend type (sqlite/noop)
  • Embedding: Provider, dimensions, embedded ratio
  • Provider: Active LLM provider and status (online/offline)
  • Chain: Fallback chain configuration
  • Tools: Number of registered tools
  • Events: Event bus activity
  • Messages: Total processed, denied count
  • LLM: Call count and average latency

​
health

Perform detailed health check across all 5 runtime layers. Usage: 
health
Example Output: 
OneClaw Health Check:
  L0 Security:     OK
  L1 Orchestrator: OK (provider: anthropic (online))
  L2 Memory:       OK (42 entries)
  L3 Event Bus:    OK (0 pending)
  L4 Tools:        OK (3 registered)

  Uptime: 2h 15m 30s | Status: HEALTHY
Security: Requires system:health authorization Status Values:
  • HEALTHY: All layers operational, LLM provider online
  • DEGRADED: Core layers OK, but LLM provider offline
  • UNHEALTHY: One or more critical layers failed
Layer Checks:
  • L0 Security: Authorization subsystem probe
  • L1 Orchestrator: LLM provider availability
  • L2 Memory: Storage backend query
  • L3 Event Bus: Pending event count
  • L4 Tools: Registry population

​
metrics

Display full operational telemetry with performance counters. Usage: 
metrics
Example Output: 
OneClaw Metrics:
  Uptime: 2h 15m 30s

  Messages:
    Total: 87
    Secured: 85
    Denied: 2
    Rate Limited: 0

  LLM:
    Calls: 15
    Failed: 1
    Tokens: 12,450
    Latency (total): 3,675ms
    Latency (avg): 245ms

  Memory:
    Stores: 42
    Searches: 28

  Tools:
    Calls: 8
    Failed: 0

  Events:
    Processed: 1,250

  Errors: 1
Security: Requires system:metrics authorization Metric Categories:
  • Messages: Total processed, security outcomes, rate limiting
  • LLM: API calls, failures, token usage, latency
  • Memory: Store/search operations
  • Tools: Execution count and failures
  • Events: Event bus processing
  • Errors: Total error count

​
reload

Check configuration file for changes and report differences. Usage: 
reload
Example Output: 
Config reload from: config/oneclaw.toml
  providers.default: ollama -> anthropic
  providers.anthropic.model: claude-sonnet-4 -> claude-opus-4
  (Report only — restart to apply changes)
Security: Requires system:reload authorization Behavior:
  • Searches for oneclaw.toml or config/oneclaw.toml
  • Parses new config and compares to current runtime config
  • Reports differences across key fields:
    • providers.default
    • security.deny_by_default
    • memory.backend
    • runtime.name
    • providers.*.model
    • channels.active
  • Does NOT hot-apply changes — requires restart
No Changes: 
Config reload from: config/oneclaw.toml
  No changes detected.

​
exit / quit

Gracefully shut down the OneClaw runtime. Usage: 
exit
quit
q
Output: 
Goodbye!
Graceful shutdown initiated
Event loop stopped.
Security: Always open (no authorization required) Behavior:
  • Sends “Goodbye!” response to channel
  • Drains event bus queue
  • Stops event loop
  • Clean shutdown (no force-kill)
Variants: All case-insensitive (EXIT, Quit, etc.)

Build docs developers (and LLMs) love

Get started for freeTalk to us