cve-remediation

Pass

Audited by Gen Agent Trust Hub on Jul 19, 2026

Risk Level: SAFECOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
  • Command Execution: The skill utilizes standard project build tools and wrappers, such as Maven (./mvnw), Gradle (./gradlew), and the .NET CLI, to apply security patches and verify build integrity. These commands are essential for the skill's primary function of dependency remediation and represent standard developer operations.
  • Indirect Prompt Injection: This skill ingests and parses JSON findings from external vulnerability reports (cve-report-N.json). This process involves the following security considerations:
  • Ingestion points: Vulnerability data is read from JSON reports generated by tools or fallback scanning in SKILL.md (Step 3).
  • Boundary markers: Explicit delimiters or specific warnings to ignore embedded instructions within the ingested JSON content are not defined.
  • Capability inventory: The skill has the capability to execute shell commands and modify manifest files, as detailed in SKILL.md Steps 4 and 6.
  • Sanitization: The instructions do not specify a validation or sanitization step for the text within vulnerability explanations before the agent processes them.
Audit Metadata
Risk Level
SAFE
Analyzed
Jul 19, 2026, 05:02 PM
Security Audit — agent-trust-hub — cve-remediation