arkcli-code-example
Warn
Audited by Gen Agent Trust Hub on Oct 17, 2026
Risk Level: MEDIUMCREDENTIALS_UNSAFECOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
- [CREDENTIALS_UNSAFE]: The skill documentation and internal logic (via arkcli) state that the placeholder '$ARK_API_KEY' in generated examples is automatically replaced with the user's actual API key from their active profile. This results in sensitive credentials being written in plaintext into various source code files (e.g., .py, .go, .sh) on the local filesystem, which increases the risk of accidental exposure if these files are shared or committed to version control.
- [INDIRECT_PROMPT_INJECTION]: The skill exhibits an indirect injection surface by fetching code blocks from an external data source (Volcengine OpenGetSampleCode API) and writing them directly to the local filesystem. While it uses the 'arkcli' tool, the skill allows user-defined output directories ('--output-dir'), which could be exploited to write content to unintended or sensitive locations if path validation is insufficient.
- [COMMAND_EXECUTION]: The skill relies on executing the 'arkcli' binary to interact with the Volcengine platform, perform network requests to official endpoints, and manage local file operations. This is the primary mechanism for the skill's functionality.
Audit Metadata