spring-ai
Pass
Audited by Gen Agent Trust Hub on Sep 15, 2026
Risk Level: SAFEEXTERNAL_DOWNLOADSINDIRECT_PROMPT_INJECTION
Full Analysis
- [EXTERNAL_DOWNLOADS]: The skill references documentation and update notes from official Spring Project domains (docs.spring.io). These are well-known, trusted resources for Java development and do not involve executable code downloads.
- [INDIRECT_PROMPT_INJECTION]: The skill has an ingestion surface as it processes external project files including pom.xml, build.gradle, and Java source files through its diagnostic script. The script maps findings to a structured JSON format to provide clear boundaries for the agent. While it processes untrusted local data, it includes explicit redaction logic to ensure detected secrets or credentials are not included in the output passed to the agent.
- Ingestion points: pom.xml, build.gradle, application.properties, and .java files (scripts/spring_ai_check.py)
- Boundary markers: Structured JSON output (SKILL.md)
- Capability inventory: Local file reading only; no network or subprocess capabilities (scripts/spring_ai_check.py)
- Sanitization: Automated redaction of literal API keys and secrets (scripts/spring_ai_check.py)
- [SAFE]: The core logic is implemented in a standalone Python script that uses only the standard library. The script is designed for read-only static analysis and does not perform network operations, spawn subprocesses, or modify the filesystem, aligning with security best practices for diagnostic tools.
Audit Metadata