linkpix-image-generate

Fail

Audited by Gen Agent Trust Hub on Sep 21, 2026

Risk Level: HIGHPERSISTENCEEXTERNAL_DOWNLOADSCOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
  • [PERSISTENCE]: The skill instructions include a step to append an 'export PATH' command to the user's '~/.bashrc' file. This is intended to ensure that the newly installed Node.js binaries are available in future shell sessions. While this is a common practice for software installation, it constitutes a persistence mechanism that modifies shell configuration files.
  • [EXTERNAL_DOWNLOADS]: The skill downloads Node.js binary distributions from 'nodejs.org' and the '@iqinghu/qhkit' package from the official NPM registry. It also provides 'registry.npmmirror.com' (Alibaba mirror) as a fallback. These are well-known and reputable services for software distribution. The skill correctly implements integrity verification using 'sha256sum' for the downloaded binaries.
  • [COMMAND_EXECUTION]: The skill uses several shell commands to manage its environment and perform tasks, including 'curl', 'tar', 'npm', and 'qhkit'. It includes a verification command 'curl ... | grep ... | sha256sum -c -' which is used for security validation of the downloaded archives, rather than direct execution of remote scripts. The skill also executes the 'qhkit' CLI tool to generate images based on user input.
  • [INDIRECT_PROMPT_INJECTION]: The skill processes untrusted user data via the 'prompt' and 'uploadedImages' parameters, which are then passed to the 'qhkit' command-line tool. This creates a potential surface for indirect prompt injection attacks.
  • Ingestion points: The 'prompt' (textual description) and 'uploadedImages' (local file paths or URLs) fields defined in the 'SKILL.md' recipes.
  • Boundary markers: The instructions do not specify any delimiters or safety markers to differentiate user-provided data from the agent's internal instructions when calling the tool.
  • Capability inventory: The skill can execute shell commands ('qhkit', 'npm'), perform file system operations ('tar', 'mkdir'), and access the network ('curl', 'npm').
  • Sanitization: There is no mention of sanitization, validation, or escaping logic for the user-supplied content before it is processed by the external tool.
Recommendations
  • HIGH: Downloads and executes remote code from: https://nodejs.org/dist/v22.22.3/SHASUMS256.txt - DO NOT USE without thorough review
Audit Metadata
Risk Level
HIGH
Analyzed
Sep 21, 2026, 04:09 PM