ai-context-exclusion
Pass
Audited by Gen Agent Trust Hub on Nov 5, 2026
Risk Level: SAFE
Full Analysis
- [DATA_EXPOSURE_AND_EXFILTRATION]: The skill proactively identifies sensitive file paths and patterns—such as
.env,google-services.json,credentials.*, andkeyboards—to ensure they are added to exclusion lists. This is a security best practice designed to prevent accidental data exposure to AI models. - [INDIRECT_PROMPT_INJECTION]: The skill reads existing configuration files (like
.claude/settings.json) to perform 'safe merges.' While it processes local files, the instructions focus on maintaining existing security parameters and adding new exclusion rules, which reduces the overall risk of prompt injection by limiting what the AI can see. - [SAFE]: The skill performs legitimate project configuration tasks. It does not use any network tools, execute arbitrary code, or attempt to bypass safety guidelines. Its purpose is entirely defensive.
Audit Metadata