Audit History

All eight static findings are false positives. The six 'external_commands' hits are the static analyzer misclassifying Markdown code-fence markers and documentation table rows as Ruby/shell backtick execution; the two 'env_access' hits are declarative references to the NANSEN_API_KEY environment variable name in the skill's metadata (requires/primaryEnv), not code that reads or exfiltrates the secret. The skill only documents invocations of a first-party 'nansen' CLI with hardcoded subcommands and no command-injection vector. No prompt injection, data exfiltration, or business-logic abuse was found.

The Markdown backtick findings are false positives; they are code fences and inline flag formatting. The NANSEN_API_KEY findings are confirmed because the skill requires a secret-bearing environment variable for the Nansen CLI, with no evidence of prompt injection or data exfiltration in SKILL.md.