review-architecture

Pass

Audited by Gen Agent Trust Hub on Jun 19, 2026

Risk Level: SAFEPROMPT_INJECTION
Full Analysis
  • [SAFE]: The skill is entirely instructional and does not contain any executable scripts, binary files, or hidden code.
  • [PROMPT_INJECTION]: The skill is designed to process untrusted code diffs, which introduces an indirect prompt injection surface. A malicious diff could contain embedded instructions intended to manipulate the reviewer's output or bias the architectural report. However, the skill's restricted output format and read-only nature mitigate potential impact.
  • Ingestion points: The diff input parameter used during the review phase as defined in SKILL.md.
  • Boundary markers: The instructions do not specify any delimiters or explicit 'ignore embedded instructions' warnings for the input diff content.
  • Capability inventory: The skill is limited to generating a text-based report and a verdict; it has no access to the file system, network, or shell command execution.
  • Sanitization: No sanitization or validation of the input diff is specified in the prompt instructions.
Audit Metadata
Risk Level
SAFE
Analyzed
Jun 19, 2026, 08:34 PM
Security Audit — agent-trust-hub — review-architecture