Operator Commands

Day-to-day operation of WASP happens through four surfaces: the wasp CLI, natural language via Telegram, the dashboard, and direct Docker commands. This page is the cheat-sheet.

wasp CLI

Installed at /usr/local/bin/wasp by install.sh. Run any command from any directory.

wasp onboard         Re-run the configuration wizard
wasp start           Start the stack
wasp stop            Stop the stack
wasp restart         Restart all services
wasp status          Container status + health checks
wasp logs [service]  Stream logs (default: agent-core)
wasp health          Run the full health probe suite
wasp update          Pull latest tarball, rebuild, restart, verify
wasp backup          Create a timestamped backup archive (Postgres + volumes)
wasp restore <file>  Restore from a backup archive
wasp reset           Wipe runtime state but keep volumes
wasp uninstall       Remove WASP (asks before deleting data)
wasp help            Show command reference

Telegram

WASP listens on the bot configured by TELEGRAM_BOT_TOKEN. Only Telegram user IDs in TELEGRAM_ALLOWED_USERS can interact.

Asking the agent something

Send any message in your natural language. The agent will:

1. Try a deterministic fast-path (e.g., listing reminders, deleting a reminder, listing tasks).
2. Run the Decision Layer to classify intent (5 strategies: DIRECT_RESPONSE / GOAL / SCHEDULED_TASK / SUB_AGENT / SCRIPT).
3. If the request needs planning, create a Goal; if it is a one-shot answer, run the LLM loop directly.
4. Apply the policy guards before sending the reply.

Built-in Telegram commands

Command	Effect
/start	Concise welcome (EN/ES/PT/FR) with capability summary
/help	Full command reference
/ping	Reachability check
/status	System status
/model	Show active model + provider
/skills	List registered skills
/skill <name>	Invoke a skill directly
/schedule	Show scheduled tasks
/memory	Memory subsystem info
/snapshot	Save current state to a memory snapshot
/introspect	Capability + health snapshot
/monitor <url>	Watch a URL for changes
/broker	Integrations management
/api set <provider> <key>	Persist a model API key
/openclaw <action>	Manage dynamic skills from the ClawHub registry

Telegram input types

Input	Handling
Text	Full pipeline
Photo	Vision-capable model receives the image bytes
Document	Treated as file; can be referenced in subsequent messages
Voice note	Transcribed via OpenAI Whisper, then passed through the text pipeline
Video / video note	First frame extracted with ffmpeg, then vision pipeline

Live progress

While the agent is working on a response, it edits a single status message in place (TELEGRAM_PROGRESS events) showing the current step. Only one progress message is created per turn; on edit failure, it is dropped silently.

Common natural-language workflows

Create a recurring task

every 6 hours, fetch BTC price and email me a report

The agent detects the recurring keyword, routes to task_manager, creates a custom task with the literal instruction. The Custom Task Runner job (60 s) triggers it at the next interval.

:::warning Schedule semantics task_manager only supports interval scheduling. Clock-time requests (every Monday at 9am) and daypart phrases (in the morning) are surfaced via an automatic disclaimer; they are NOT honored literally. To approximate a fixed clock time, create the task at the desired wall-clock time so the interval boundary aligns. :::

Create a reminder

remind me to back up the disk in 3 hours

Reminders are one-shot or recurring. The Reminder Checker (30 s) fires due reminders. If a reminder is linked to an agent, firing it restarts that agent's goal cycle.

Create a sub-agent

create an agent named News Watcher to monitor RSS feeds every hour

The handler extracts the name (non-greedy regex with lookahead stop-set), the purpose, and the autonomy mode. Sub-agents have their own goal queues and Telegram chat-id default. Manage them at /agents.

One-time screenshot

capture a screenshot of example.com

The agent runs browser(action="capture", url="..."). The screenshot arrives as a Telegram photo with the page title as caption.

Investigate logs

there are 503 errors in the audit log, find a pattern and propose a fix

The agent creates a Goal: query audit log → analyze → propose patch via self_improve(action="propose"). Review the proposal at /self-improve. Apply only after reading the diff.

Host shell

Action	Command
Container status	docker compose ps
Recent agent logs	docker compose logs agent-core --tail=200
Follow logs	docker compose logs -f agent-core
Health check	curl -s http://localhost:8080/health | jq
Queue depths	docker exec agent-redis redis-cli XLEN events:incoming
Behavioral queue	docker exec agent-redis redis-cli LLEN behavioral:pending
Postgres console	docker exec -it agent-postgres psql -U agent -d agent
Redis console	docker exec -it agent-redis redis-cli
Restart core	docker compose up -d agent-core
Rebuild core	docker compose build agent-core && docker compose up -d agent-core

Dashboard quick reference

Page	Use it for
/overview	Top-level snapshot
/health	CPU pressure (CPI), memory, queues, model liveness
/traces	Per-response forensic trace
/audit	Per-action AuditLog (keyset pagination)
/cmd	Command palette (search across pages)
/scheduler	Recurring tasks, custom tasks, monitor list
/agents	Sub-agent CRUD, status, message log
/goals	Active and historical goals with TaskGraph visualization
/self-improve	Pending AI proposals — diff viewer, Apply/Reject
/behavioral-rules	Learned rules — view, filter, toggle, delete
/config	prime.md editor + 12 feature flags
/models	Provider status, default model, catalog
/integrations	40+ connectors with circuit-breaker state
/cognitive	Self-model, epistemic state, integrity report
/knowledge-graph	Force-directed canvas of entities and relations
/world-model	Temporal observations, entity timelines
/memory	Memory tree browser
/vector-memory	Semantic similarity search
/reset	Panic Reset (hard confirmation)

See also

• Telegram Integration — full input handling reference
• Dashboard Integration — every dashboard page in detail
• Configuration — prime.md and feature flags
• Monitoring — health and metrics
• Logs — what to look for