solana-sniper-bot

MEDIUM W011: Third-party content exposure detected (indirect prompt injection risk).

• Third-party content exposure detected (medium risk: 0.30). The workflow monitors public Raydium pools and passes token metadata and pool details into an LLM risk assessment prompt without pre-filtering or sanitizing the untrusted data.

MEDIUM W012: Unverifiable external dependency detected (runtime URL that controls agent).

• Potentially malicious external URL detected (high risk: 0.70). The skill makes programmatic HTTP calls at runtime to external third-party endpoints (Raydium, Jupiter, Anthropic, Solana) to fetch data and execute trades, representing unverified runtime external operational dependencies.

MEDIUM W009: Direct money access capability detected (payment gateways, crypto, banking).

• Direct money access detected (high risk: 0.70). The skill operates an autonomous Solana trading bot that executes real-money purchases and trades of cryptocurrency tokens using a private key (SOLANA_PRIVATE_KEY) and an aggregator (Jupiter/Raydium). This is a direct financial operation (trades/purchases). No independent service-layer access control or authentication check is documented for the trading route other than supplying the private key and executing transactions.

MEDIUM W013: Attempt to modify system services in skill instructions.

• Attempt to modify system services in skill instructions detected (high risk: 0.80). The skill instructs the agent to copy script files into system directories like /opt/sniper/ which typically requires administrative/sudo privileges, potentially modifying system-wide directories.