suggest-improvements
Pass
Audited by Gen Agent Trust Hub on Sep 19, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONCOMMAND_EXECUTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill is designed to read and analyze untrusted external data from a user's project files, including
README.mdandpackage.json. Malicious instructions embedded in these files could potentially influence the agent's typography recommendations or subsequent actions. - Ingestion points: Reads local project files (
package.json,README.md, style files). - Boundary markers: The instructions do not define specific delimiters or "ignore instructions" warnings for the file analysis phase.
- Capability inventory: The skill utilizes file read access, executes MCP tools (
search_pairings,preview_pairing,install_pairing), and suggests terminal commands. - Sanitization: No explicit validation or filtering of the content from project files is mentioned before processing.
- [COMMAND_EXECUTION]: The skill recommends the execution of a shell command to install font pairings:
bunx shadcn@latest add https://www.fonttrio.xyz/r/{pairing-name}.json. This pattern involves fetching and applying remote configuration from a service URL to the user's local environment.
Audit Metadata