gmgn-holder-analysis

HIGH W007: Insecure credential handling detected in skill instructions.

• Insecure credential handling detected (high risk: 0.80). The skill instructs running and showing gmgn-cli config and to apply a key via --apply <KEY>, and it requires pasting the tool's complete stdout verbatim — forcing the agent to handle and potentially output secret/API key values verbatim.

CRITICAL E006: Malicious code pattern detected in skill scripts.

• Malicious code pattern detected (high risk: 1.00). The skill contains explicit instructions that encourage revealing raw CLI output and running a config flow (including "show output" and applying keys), and the code funnels all data through an external CLI (gmgn-cli) while requiring verbatim pasting of stdout — this creates a clear data-exfiltration/credential-leakage risk vector.

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

• Third-party content exposure detected (high risk: 0.75). The required runtime path runs python3 .../analyze.py, which calls gmgn-cli token holders and gmgn-cli portfolio created-tokens; those CLI results (public on-chain-derived holder/dev/token history data returned by the vendor) become readable JSON fields and are embedded into the agent’s LLM context via the “paste stdout verbatim” output rule—i.e., untrusted external data from non-user/outsider sources.