nonce-manager
Pass
Audited by Gen Agent Trust Hub on Oct 1, 2026
Risk Level: SAFECOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill documents and executes its own CLI subcommands (acquire, release, sync, status) using the Bun runtime to manage transaction nonces.\n- [INDIRECT_PROMPT_INJECTION]: The agent processes JSON output from CLI commands, creating an indirect prompt injection surface.\n
- Ingestion points: The agent reads output from the acquire, sync, and status commands in nonce-manager.ts.\n
- Boundary markers: Absent; no specific delimiters or validation logic are shown for the tool's output within the agent context.\n
- Capability inventory: The skill is capable of file system writes to ~/.aibtc/nonce-state.json and network requests to the Hiro API for data synchronization.\n
- Sanitization: Absent; the tool returns raw state information and blockchain data without explicit sanitization for LLM consumption.
Audit Metadata