kotlin-multiplatform-shadcn-compose
Pass
Audited by Gen Agent Trust Hub on Aug 1, 2026
Risk Level: SAFECOMMAND_EXECUTIONEXTERNAL_DOWNLOADSPROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill includes a Python script
scripts/fetch_component_signature.pythat the agent is instructed to run. This script is used to verify component signatures against the library's live source code to prevent hallucination of parameters. - [EXTERNAL_DOWNLOADS]: The helper script
scripts/fetch_component_signature.pyperforms network requests toapi.github.comandraw.githubusercontent.comto retrieve repository metadata and Kotlin source files for analysis. Additionally, the skill guides the user to add theio.github.ronjunevaldoz:shadcn-composedependency via Maven Central. - [PROMPT_INJECTION]: The skill exhibits an indirect prompt injection surface by ingesting untrusted data from a remote repository via the
fetch_component_signature.pyscript. This data is used by the agent to generate code for the user. - Ingestion points:
api.github.comandraw.githubusercontent.com(fetching Kotlin source code). - Boundary markers: Absent. The script output is trusted directly for code generation.
- Capability inventory: Subprocess execution (running the fetch script) and code generation within the agent's context.
- Sanitization: Absent; the script extracts function signatures using regex and balanced-parentheses scanning without validating the content for malicious instructions.
Audit Metadata