The Agent Skills Directory

[SAFE]: The skill consists entirely of markdown documentation and reference materials for configuring the Zed editor. No malicious code, obfuscation, or unauthorized access patterns were detected. The skill specifically notes that API keys should be stored in the OS keychain rather than configuration files.
[INDIRECT_PROMPT_INJECTION]: The skill facilitates the management of Zed's rules system, which involves processing instructions from untrusted repository files like .rules or .cursorrules.
Ingestion points: Reference files listed in references/rules.md (e.g., .rules, .cursorrules, .windsurfrules).
Boundary markers: No explicit boundary markers or isolation instructions for external rules are defined in the skill documentation.
Capability inventory: The skill manages settings that control tool permissions and can modify editor behavior, but it focuses on standard editor configuration.
Sanitization: The skill provides reference examples for defensive patterns (e.g., always_deny regex for dangerous commands), helping users sanitize their own environment permissions.
Assessment: This is a low-risk finding as it describes standard editor functionality and provides guidance on how to configure it securely.

zed-config