Skip to main content
Start the interactive shell with opensre (TTY required). Type a slash command at the prompt, or describe what you want in plain language and the planner can map intent to the right command. Run /help anytime for the full, up-to-date command list grouped by category.

​
Quick reference

CommandWhat it does
/helpList commands or show help for one command (/help /model, /help tasks)
/statusSession summary — interactions, alerts, provider, reasoning effort, trust mode
/costToken usage and LLM call count for the current session
/effortSet or show reasoning effort (low, medium, high, xhigh, max)
/clearClear the screen and re-render the banner
/exitLeave the interactive shell (/quit is an alias)
/healthIntegration and agent health
/investigateRun an investigation from the REPL
/integrationsList, verify, or manage integrations
/modelShow or switch LLM provider and models
/sessionsList recent REPL sessions
/resume <id>Restore a past session’s conversation context
For session persistence in depth, see Session History. For /history and /privacy, see Interactive Shell Privacy.

​
Session cost — /cost

Show how much LLM usage the current REPL session has accumulated: 
/cost
Typical output includes:
FieldMeaning
history entriesLines recorded in this session’s history
llm callsNumber of LLM turns counted this session
input tokens / output tokensTotals for prompt and completion usage
When the provider returns usage metadata (for example on planner invoke calls), those counts are measured. Streaming chat and help paths estimate tokens from text length (roughly four characters per token). If any estimates are included, the table title notes (includes estimates) and measured vs estimated splits appear on the token rows. Token totals reset when you start a new session (/new) or clear session state; /cost does not show billing from your cloud provider — only what OpenSRE tracked locally for this shell session.

​
Session status — /status

/status
Shows interaction count, incoming alerts, last investigation, trust mode, reasoning effort, active provider, and any accumulated infra context keys.

​
Reasoning effort — /effort

For OpenAI and Codex providers, set how much reasoning the model applies in this REPL session: 
/effort high
/effort
Bare /effort prints the current level and usage. /status includes the same field. Other providers ignore this setting. You can also set OPENSRE_REASONING_EFFORT in the environment for non-interactive defaults. See LLM providers.

​
Help — /help

/help
/help /model
/help investigation
/help all
In a TTY, bare /help opens an interactive picker. Use /help <command> or /help <category> for details on one topic.