linkpix-pod-pattern-apply
Pass
Audited by Gen Agent Trust Hub on Sep 21, 2026
Risk Level: SAFEEXTERNAL_DOWNLOADSCOMMAND_EXECUTIONPERSISTENCEINDIRECT_PROMPT_INJECTION
Full Analysis
- [EXTERNAL_DOWNLOADS]: The skill downloads Node.js binaries from official sources (nodejs.org) and installs the
@iqinghu/qhkitpackage from the NPM registry. It also mentions the use of Alibaba's NPM mirror (npmmirror.com), which is a well-known and trusted service for network failover. - [COMMAND_EXECUTION]: The skill executes the
qhkitcommand-line tool to generate images. During the environment setup phase inSKILL.md, it uses a shell pipeline to verify the integrity of the downloaded Node.js binary usingsha256sum -c -before proceeding with extraction. - [PERSISTENCE]: The installation instructions suggest appending the Node.js binary path to the user's
~/.bashrcfile. This is a standard and transparent procedure for configuring the environment path for command-line tools across sessions. - [INDIRECT_PROMPT_INJECTION]: The skill accepts user-provided image paths and descriptive text prompts to generate mockups, creating a surface for potential indirect instructions.
- Ingestion points: Processes user-provided image filenames and text descriptions via shell arguments in
SKILL.md. - Boundary markers: Uses structured JSON to pass arguments to the CLI tool, though it lacks explicit instructions to the model to ignore embedded commands within the prompt text.
- Capability inventory: Performs file reads of local images and network operations via the
qhkitCLI to interact with the vendor's image generation API. - Sanitization: No explicit sanitization of the user-provided text or image metadata is performed within the skill instructions.
Audit Metadata