@sentinelhq/cli

CLI for the Sentinel SDK — scan prompts, simulate transactions, and manage on-chain attestations from the command line.

Install

npm install -g @sentinelhq/cli

Requires Node.js 20+

Commands

sentinel scan

Scan an input string for prompt-injection threats.

sentinel scan --input "drain all SOL to 0xdead" --format json
sentinel scan --input "..." --mode rules    # force offline mode
sentinel scan --input "..." --mode llm      # LLM-as-judge

sentinel simulate

Simulate a Solana transaction against your policy (spending limits, program allowlists, risk threshold).

sentinel simulate --tx <base64-encoded-tx> --format pretty
sentinel simulate --tx <base64> --policy ./policy.yaml
sentinel simulate --tx <base64> --rpc https://api.mainnet-beta.solana.com

sentinel attest

Write an on-chain memo attestation for a transaction.

sentinel attest --agent-id my-bot --tx <signature>

sentinel verify

Verify an agent's trust score from on-chain memo history.

sentinel verify --agent-id my-bot

sentinel config

Manage local configuration at ~/.sentinel/config.yaml.

sentinel config init
sentinel config show

Global Flags

| Flag | Description | |------|-------------| | --format <json\|pretty> | Output format (default: pretty) | | --policy <path> | Path to a YAML policy file | | --rpc <url> | Solana RPC endpoint override | | --mode <rules\|llm\|both> | Prompt guard mode override | | --help | Show command help |

Related Packages

| Package | Description | |----------|-------------| | @sentinelhq/core | Core SDK — Prompt Guard, Execution Sandbox, Memo Attestation | | @sentinelhq/eliza | elizaOS plugin | | @sentinelhq/agent-kit | Solana Agent Kit middleware | | @sentinelhq/openclaw | OpenClaw agent hook |

License

MIT