The Agent Skills Directory

[EXTERNAL_DOWNLOADS]: The skill instructs the agent to install a third-party dependency huashu-design from an external source (alchaincyf/huashu-design) that is not identified as a well-known or trusted vendor resource.
[REMOTE_CODE_EXECUTION]: The installation of the external dependency is performed via npx skills add alchaincyf/huashu-design -g -y, which executes remote package code on the host system during the setup process.
[COMMAND_EXECUTION]: The skill requires the execution of shell commands such as test -f to validate paths and assumes the ability to run verification scripts or browser-based tools like Playwright to confirm the output.
[PROMPT_INJECTION]: The skill is subject to indirect prompt injection as it ingests untrusted external data (brand guidelines, design assets, and URLs) to generate high-fidelity prototypes and reviews.
Ingestion points: Design assets, brand guidelines (files/URLs), and product screenshots processed in SKILL.md (Step 2).
Boundary markers: Absent. The instructions do not specify any delimiters or warnings to the agent regarding embedded instructions in user-provided assets or external URLs.
Capability inventory: The skill possesses extensive file-write capabilities (across temp, state, and delivery directories) and shell execution permissions through its automated workflow.
Sanitization: Absent. No validation or filtering of the content from external inputs is performed before the agent processes them as part of the design creation.

soia-design-explorer