laravel-best-practices
Pass
Audited by Gen Agent Trust Hub on Nov 20, 2026
Risk Level: SAFECOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill requires the agent to execute a bundled PHP script (
scripts/laravel-profile.php) to analyze the project environment and dependencies. - [INDIRECT_PROMPT_INJECTION]: The profiling script processes data from project configuration files such as
composer.jsonandcomposer.lock. If these files are maliciously crafted (e.g., in a compromised repository), they could contain metadata intended to influence the agent's behavior when it later reads the generated profile. - Ingestion points:
composer.json,composer.lock, and file headers of PHP files within the application'sapp/directory. - Boundary markers: The script writes output to a structured Markdown file (
.agents/laravel-project-profile.md) using standard delimiters. - Capability inventory: The script performs local file read/write operations; it has no network access. The agent reads the resulting profile to inform its code generation and review tasks.
- Sanitization: The script extracts package names and version strings from JSON files and interpolates them into the Markdown profile without explicit sanitization or filtering for prompt injection patterns.
Audit Metadata