gmgn-swap

Fail

Audited by Snyk on Nov 14, 2026

Risk Level: CRITICAL
Full Analysis

CRITICAL E004: Prompt injection detected in skill instructions.

  • Potential prompt injection detected (medium risk: 0.30). The skill is a legitimate trading tool whose risks are disclosed, but its fee handling forwards unbounded fee values to the backend without client-side limits, which weakens protection on the user's funds beyond what the task needs, so this is a careless risk rather than malicious intent.

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

  • Third-party content exposure detected (low risk: 0.10). The required workflow reads on-chain token data from GMGN's API, including token names in order strategy list responses (token_name, line 658), which anyone can set when deploying a token. An outsider can publish such a token but cannot force the agent to read it, so this is low rather than medium.

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

  • Direct money access detected (medium risk: 0.30). Direct operation: the skill submits real, irreversible on-chain token swaps, limit and stop orders, and multi-wallet batch trades (SKILL.md:61-63, 131-176, 315-357). It is not an indirect or excluded case. Control: the route requires GMGN_API_KEY and GMGN_PRIVATE_KEY (signed auth), and the API key is transmitted to GMGN's servers in the X-APIKEY header (SKILL.md:39, 53, 402). The service also enforces that the --from wallet and every multi-swap account must be bound to that API key (SKILL.md:183, 364). This is a service-validated credential check, which the rule counts as access control, not a local key-presence check or agent-followed instruction. The local signing alone would not qualify, but the server-side API key validation and wallet-binding enforcement do. Medium rather than high.

Issues (3)

E004
CRITICAL

Prompt injection detected in skill instructions.

W011
MEDIUM

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

W009
MEDIUM

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

Audit Metadata
Risk Level
CRITICAL
Analyzed
Nov 14, 2026, 02:47 AM
Issues
3