stack-detection

Pass

Audited by Gen Agent Trust Hub on Jul 3, 2026

Risk Level: SAFEDATA_EXFILTRATIONPROMPT_INJECTION
Full Analysis
  • [DATA_EXFILTRATION]: The skill instructs the agent to access potentially sensitive files such as .env and .env.example. This access is intended for identifying the runtime environment (PHP version, services, and extensions) as part of a security audit. No network commands (e.g., curl, wget) are present to exfiltrate this data.
  • [PROMPT_INJECTION]: The skill is susceptible to indirect prompt injection because it ingests and processes untrusted data from repository files like composer.json and Bitrix version.php.
  • Ingestion points: Files including composer.json, composer.lock, .env, and framework-specific version files.
  • Boundary markers: None identified; the instructions do not specify using delimiters or warnings when processing these files.
  • Capability inventory: Primarily file-read operations used to generate a structured "stack profile" artifact.
  • Sanitization: No explicit validation or sanitization of the content read from the repository is mentioned.
Audit Metadata
Risk Level
SAFE
Analyzed
Jul 3, 2026, 04:16 PM
Security Audit — agent-trust-hub — stack-detection