The Agent Skills Directory

[SAFE]: The skill logic is highly structured and consistent with its stated purpose as a regulatory compliance auditor. No malicious patterns, obfuscation, or unauthorized data exfiltration were detected across the instructions or reference materials.- [INDIRECT_PROMPT_INJECTION]: The skill has a surface for indirect prompt injection as it processes untrusted codebase data during its Phase 1 audit. This risk is mitigated by the skill's prescriptive instructions, its strictly read-only scanning phase, and the mandatory use of the AskUserQuestion tool to obtain explicit user confirmation before implementing any remediation fixes or installing dependencies.- [EXTERNAL_DOWNLOADS]: The skill references official regulatory and standards documentation from trusted well-known services including NIST, the European Union, and ISO. It also recommends the installation of legitimate, industry-standard security and privacy libraries such as guardrails-ai and presidio-analyzer to address identified compliance gaps.

arize-compliance-audit