The Agent Skills Directory

[SAFE]: The skill acts as a documentation and orientation index for various technical mechanisms. No malicious code, prompt injection, or suspicious obfuscation patterns were detected.
[PROMPT_INJECTION]: The skill uses natural language instructions to guide the agent through its registry. No attempts to bypass safety filters or override system instructions were detected.
[DATA_EXFILTRATION]: No hardcoded credentials or access to sensitive system files (e.g., .ssh, .env) were found. The skill's toolset is limited to read-only operations (read, grep, glob) on its own files.
[EXTERNAL_DOWNLOADS]: The skill contains references to many well-known and trusted external documentation sources, including Wikipedia, official project sites (e.g., json-schema.org, postgresql.org, sqlite.org), and Mozilla technical blogs. These links are informational and do not involve the execution of untrusted remote content.
[COMMAND_EXECUTION]: The skill metadata identifies several CLI tools (e.g., jq, sqlite3, yq, psql) associated with specific mechanisms. However, the skill itself does not execute these commands; it merely documents their roles for the agent's situational awareness.
[REMOTE_CODE_EXECUTION]: No patterns of downloading and executing remote scripts (e.g., curl | bash) were found. Installation commands for packages like pywin32 and comtypes are mentioned in documentation for informational purposes only.

schema