check-bounded-contexts

Pass

Audited by Gen Agent Trust Hub on Mar 17, 2026

Risk Level: SAFECOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
  • [COMMAND_EXECUTION]: The skill uses local system commands including grep, find, and echo within shell loops to identify bounded contexts and detect coupling patterns in the project directory.
  • [PROMPT_INJECTION]: The skill is vulnerable to indirect prompt injection as it processes external data from PHP source files and configuration files.
  • Ingestion points: PHP source code files within the src/ directory and the composer.json file.
  • Boundary markers: Absent; the skill does not use delimiters to isolate processed content from its instructions.
  • Capability inventory: The skill primarily uses read-only shell commands to extract information but lacks destructive file system or network capabilities.
  • Sanitization: No sanitization is performed on the content extracted from the source files before it is used to generate the final analysis report.
Audit Metadata
Risk Level
SAFE
Analyzed
Mar 17, 2026, 09:57 AM