php-auth-audit
Pass
Audited by Gen Agent Trust Hub on Sep 17, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill is designed to ingest and process untrusted data from external PHP source code and metadata files generated by other tools in a pipeline (e.g.,
routes_{timestamp}.md,params_{timestamp}.md). - Ingestion points: The skill reads source code from a user-provided
source_pathand consumes routing/parameter metadata from pre-defined file paths. - Boundary markers: The instructions lack specific boundary markers or 'ignore embedded instructions' delimiters to separate the audited code content from the agent's internal logic.
- Capability inventory: The skill has file-writing capabilities to generate markdown reports, audit logs, and mapping tables in the
{output_path}. It does not perform network operations or direct shell execution. - Sanitization: There are no defined sanitization, escaping, or validation steps for the content of the audited files, which could theoretically allow malicious strings in the source code (like comments or variable names) to influence the agent's behavior during the audit process.
Audit Metadata