skills/bankrbot/clawdbot-skill/yoink/Gen Agent Trust Hub

yoink

Pass

Audited by Gen Agent Trust Hub on Mar 9, 2026

Risk Level: SAFEEXTERNAL_DOWNLOADSCOMMAND_EXECUTION
Full Analysis
  • [EXTERNAL_DOWNLOADS]: The skill interacts with the official Base network RPC endpoint (https://mainnet.base.org) to fetch game state data such as the current flag holder and scores.
  • [COMMAND_EXECUTION]: Utilizes standard system utilities curl and jq to perform read-only queries against blockchain APIs, as defined in the contract interface and workflow sections.
  • [PROMPT_INJECTION]: No patterns indicative of direct prompt injection or safety guideline bypasses were identified. The skill presents an indirect prompt injection surface:
  • Ingestion points: Smart contract state data returned from the Base RPC endpoint (e.g., eth_call results).
  • Boundary markers: Absent; the skill relies on the agent to parse JSON results from the RPC call.
  • Capability inventory: Transaction execution capability through the Bankr interface (to, data, value JSON format).
  • Sanitization: None specified for the data returned from the blockchain.
  • [DATA_EXFILTRATION]: Analysis did not detect any access to sensitive local files or unauthorized network communication patterns.
Audit Metadata
Risk Level
SAFE
Analyzed
Mar 9, 2026, 11:46 PM