codepulse-vuln-verbose
Pass
Audited by Gen Agent Trust Hub on Oct 2, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONEXTERNAL_DOWNLOADS
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill analyzes a user-provided codebase, which acts as an untrusted ingestion point. It then produces an HTML report on the local filesystem that includes quoted code excerpts from this codebase. The lack of explicit sanitization requirements for these excerpts introduces a potential for indirect prompt injection or stored cross-site scripting (XSS) within the generated report. \n
- Ingestion points: The entire target codebase is used as input context (SKILL.md). \n
- Boundary markers: No delimiters or instructions to ignore embedded instructions are specified for the ingestion process. \n
- Capability inventory: The skill performs local file writes to create the HTML report (SKILL.md). \n
- Sanitization: No instructions are provided to sanitize or escape code snippets before they are included in the HTML output. \n- [EXTERNAL_DOWNLOADS]: The skill directs the agent to query external vulnerability databases for validation. \n
- The instructions specify consulting the NIST National Vulnerability Database (NVD) and the MITRE CVE Program to verify security findings (references/severity-ratings.md). These are well-known and trusted official security resources.
Audit Metadata