laraveldaily-structure-audit
Pass
Audited by Gen Agent Trust Hub on Jun 21, 2026
Risk Level: SAFEPROMPT_INJECTION
Full Analysis
- [PROMPT_INJECTION]: The skill is susceptible to indirect prompt injection because it is designed to ingest and analyze untrusted data (PHP source code) provided by the user.
- Ingestion points: The skill scans all PHP source files within a project directory (excluding vendor, node_modules, and storage).
- Boundary markers: There are no instructions or delimiters provided to the agent to treat content inside the PHP files strictly as data or to ignore any AI instructions embedded within comments or strings in those files.
- Capability inventory: The skill primarily generates a report based on the files read; it does not explicitly invoke shell commands or network operations, though the underlying agent platform may possess these tools.
- Sanitization: No sanitization or filtering of the processed file content is mentioned to prevent embedded instructions from influencing the agent's behavior during the audit.
Audit Metadata