release-review
Pass
Audited by Gen Agent Trust Hub on Sep 15, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill presents an ingestion surface for untrusted data contained within the source code or configurations it audits.
- Ingestion points: Ingests local application project data such as Swift code, Info.plist structures, and pbxproj files using the
Read,Glob, andGreptools during the review phases outlined inSKILL.md. - Boundary markers: The instructions lack specific boundary delimiters or safety warnings to ensure the agent ignores natural language prompt instructions that may be intentionally placed inside source files or configuration keys.
- Capability inventory: The skill's environment is highly constrained, with access limited solely to read-only utilities as defined by
allowed-tools: [Read, Glob, Grep]. It lacks capabilities for network interaction, file writing, or shell command execution. - Sanitization: No data sanitization, content validation, or safety escaping rules are defined to sanitize the contents of user repository files prior to analysis.
Audit Metadata