sdd-clarify

Pass

Audited by Gen Agent Trust Hub on Jul 7, 2026

Risk Level: SAFE
Full Analysis
  • [SAFE]: No malicious patterns, obfuscation, or unauthorized data exfiltration attempts were detected in the skill instructions.
  • [DATA_EXPOSURE]: The skill is instructed to gather context from the local codebase (existing modules, specs, and conventions) to avoid redundant questions. This is a standard and expected behavior for a developer-oriented tool and does not involve exfiltration to external domains.
  • [PROMPT_INJECTION]: The instructions do not contain patterns intended to bypass safety filters, override system instructions, or extract system prompts.
  • [EXTERNAL_DOWNLOADS]: The skill references Greptile, a well-known code review and analysis service, as part of the PR review stage. This reference is used for verifying code changes and is consistent with the skill's stated purpose as a development workflow tool.
Audit Metadata
Risk Level
SAFE
Analyzed
Jul 7, 2026, 04:19 AM
Security Audit — agent-trust-hub — sdd-clarify