> ## Documentation Index > Fetch the complete documentation index at: https://docs.veri.studio/llms.txt > Use this file to discover all available pages before exploring further. # Run a command in a session's sandbox > Runs one command as the sandbox's agent user (never root) in `/workspace` or a directory under it, and returns its exit code and output (stdout capped at 64 KiB, stderr at 16 KiB). The command's network traffic goes through the session's egress rules like the agent's. Allowed while the session is `ready` (409 during a run, while an approval is pending, or paused); at most 30 per minute per session (429). Recorded on the transcript as `session.exec` (argv and exit code, never env values). ## OpenAPI ````yaml /api-reference/openapi.json post /v1/sessions/{session_id}/exec openapi: 3.1.0 info: title: Veri API description: >- REST API for the Veri RL post-training platform. All requests require a Bearer API key (`vk_` prefix). license: name: '' version: 0.1.0 servers: - url: https://api.veri.studio description: Production security: [] tags: - name: Training jobs description: Create, monitor, and manage training jobs. - name: Datasets description: Upload and connect training datasets. - name: Deployments description: Serve trained models and run inference. - name: Volumes description: Persistent file storage mounted into jobs. - name: Models description: Custom model registry deployments serve from. - name: Regions description: Discover available launch regions. - name: GPU description: Live GPU availability by provider and region. - name: Code artifacts description: Upload custom training script bundles. - name: Billing description: Credit balance and transaction history. - name: API keys description: Create and revoke API keys. - name: Account description: The authenticated caller's identity. - name: Settings description: Account-level integrations (Weights & Biases). - name: Metrics description: Prometheus metrics export for your own observability stack. - name: Evaluators description: 'Evaluators: versioned scoring rules (LLM judge, code, human).' - name: Experiments description: >- Experiments: offline runs of a pinned dataset snapshot through a target, scored by pinned evaluators. - name: Annotation queues description: >- Human review: queue traces, threads and experiment items, reserve one at a time, score them and feed corrections back into datasets. - name: Monitors description: >- Monitors: evaluators scoring a sampled share of a deployment's live traffic. - name: Observability description: Agent conversations, agents and their traffic. - name: Public runs description: Unauthenticated reads of runs their owners published to Explore. paths: /v1/sessions/{session_id}/exec: post: tags: - Sessions summary: Run a command in a session's sandbox description: >- Runs one command as the sandbox's agent user (never root) in `/workspace` or a directory under it, and returns its exit code and output (stdout capped at 64 KiB, stderr at 16 KiB). The command's network traffic goes through the session's egress rules like the agent's. Allowed while the session is `ready` (409 during a run, while an approval is pending, or paused); at most 30 per minute per session (429). Recorded on the transcript as `session.exec` (argv and exit code, never env values). operationId: exec parameters: - name: session_id in: path description: Session ID required: true schema: type: string requestBody: content: application/json: schema: $ref: '#/components/schemas/SessionExecRequest' required: true responses: '200': description: The command ran (any exit code) content: application/json: schema: $ref: '#/components/schemas/SessionExecResponse' '400': description: >- Invalid argv, cwd outside /workspace, bad env or timeout, or session ops disabled '404': description: Not found in this workspace '409': description: The session is not ready '422': description: >- Unknown field in the body (`root` / `harness_env` are refused: exec never runs as root) '429': description: More than 30 execs in the last minute on this session '501': description: The sandbox runtime does not support exec security: - bearerAuth: [] components: schemas: SessionExecRequest: type: object description: |- Unknown fields are refused (422): the runtime's operator-only switches (`root`, `harness_env`) are not part of this API, and a body naming them must fail loudly rather than run as the agent user without saying so. required: - argv properties: argv: type: array items: type: string description: |- The command and its arguments (no shell: use `["sh", "-c", "..."]` for pipes). 1..=256 entries. cwd: type: - string - 'null' description: Working directory, `/workspace` (default) or a path under it. env: type: object description: |- Extra environment variables, on top of the session's own (what the agent sees, model credentials excluded). `ANTHROPIC_*`, `OPENAI_*`, `VERI_*`, `PATH` and `LD_*` are dropped (listed in `env_dropped`). additionalProperties: type: string propertyNames: type: string timeout_s: type: - integer - 'null' format: int32 description: Seconds before the command is killed, 1..=300 (default 60). minimum: 0 additionalProperties: false SessionExecResponse: type: object required: - object - exit_code - stdout - stderr - truncated - timed_out - duration_ms - env_dropped properties: object: type: string exit_code: type: integer format: int64 stdout: type: string description: At most 64 KiB (`truncated` when cut). stderr: type: string description: At most 16 KiB (`truncated` when cut). truncated: type: boolean timed_out: type: boolean description: The command ran past `timeout_s` and was killed. duration_ms: type: integer format: int64 env_dropped: type: array items: type: string description: '`env` keys that were dropped (reserved names).' securitySchemes: bearerAuth: type: http scheme: bearer bearerFormat: API key description: API key with the `vk_` prefix. Create one from the dashboard. ```` This documentation is built and hosted on [Mintlify](https://mintlify.com), a developer documentation platform.