kotlin-api-reference
Warn
Audited by Gen Agent Trust Hub on Aug 30, 2026
Risk Level: MEDIUMCOMMAND_EXECUTIONDYNAMIC_EXECUTIONINDIRECT_PROMPT_INJECTIONEXTERNAL_DOWNLOADS
Full Analysis
- [COMMAND_EXECUTION]: The skill instructions in
SKILL.md(Steps 6 and 7) direct the agent to compile and run "minimal probes" or "behavioral tests" against the project's dependencies to verify API behavior at runtime. This involves executing arbitrary shell commands and code in the project environment.\n- [DYNAMIC_EXECUTION]: The helper scriptscripts/inspect-project.pyuses the standardxml.etree.ElementTreelibrary to parse Mavenpom.xmlfiles. This library is not secure against maliciously crafted XML data and is susceptible to XML External Entity (XXE) attacks, which could be used to read local files if a malicious build file is present in the project root.\n- [EXTERNAL_DOWNLOADS]: The skill fetches API documentation and configuration from official Kotlin and JetBrains sources, includingkotlinlang.organd official GitHub repositories for the Kotlin ecosystem. These are well-known and trusted services.\n- [INDIRECT_PROMPT_INJECTION]:\n - Ingestion points: The skill ingests untrusted data from project build files (
pom.xml,build.gradle.kts,libs.versions.toml) via theinspect-project.pyscript.\n - Boundary markers: The instructions do not specify the use of delimiters or warnings when the agent processes the findings from these files.\n
- Capability inventory: The skill can execute shell commands for compilation and testing, and performs file-write operations to generate evidence reports.\n
- Sanitization: While the Python script uses structured parsing to extract coordinates, it does not explicitly sanitize the data before it is interpolated into the agent's prompt or the final report.
Audit Metadata