hodlmm-signal-allocator

Pass

Audited by Gen Agent Trust Hub on Oct 1, 2026

Risk Level: SAFECREDENTIALS_UNSAFEEXTERNAL_DOWNLOADSCOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTIONDYNAMIC_EXECUTION
Full Analysis
  • [CREDENTIALS_UNSAFE]: The skill accesses sensitive wallet keystores and configuration files located in the ~/.aibtc/ directory to retrieve mnemonics and private keys for transaction signing. This is consistent with the skill's purpose as a DeFi allocator.
  • [EXTERNAL_DOWNLOADS]: Fetches operational data from multiple external sources, including market signals from aibtc.news, HODLMM pool data from bitflowapis.finance, blockchain state from hiro.so, and a risk index from quantum-power-map.p-d07.workers.dev.
  • [COMMAND_EXECUTION]: Broadcasts financial transactions to the Stacks mainnet and maintains a local execution log and cooldown state in ~/.hodlmm-signal-allocator-state.json.
  • [INDIRECT_PROMPT_INJECTION]: The skill represents an attack surface for indirect injection by ingesting external data:
  • Ingestion points: Signal data from aibtc.news and risk indices from quantum-power-map.p-d07.workers.dev in hodlmm-signal-allocator.ts.
  • Boundary markers: No explicit boundary markers or 'ignore' instructions are used when interpolating signal headlines into the agent context.
  • Capability inventory: Transaction signing/broadcasting and local file system writes in hodlmm-signal-allocator.ts.
  • Sanitization: Logical gates use numeric scores which are validated, but signal headlines are processed without sanitization.
  • [DYNAMIC_EXECUTION]: Utilizes dynamic import() calls to load cryptographic and blockchain libraries (@stacks/* and @bitflowlabs/core-sdk) at runtime, which is a standard pattern for managing optional dependencies in this environment.
Audit Metadata
Risk Level
SAFE
Analyzed
Oct 1, 2026, 05:03 AM