The Agent Skills Directory

[SAFE]: The skill consists entirely of documentation and instructional templates. No executable code, suspicious patterns, or malicious instructions were found.
[DATA_EXPOSURE]: The skill includes explicit instructions for agents and developers to avoid committing sensitive information, specifically noting that .env files and secrets should never be committed.
[PROMPT_INJECTION]: The skill provides defensive guidance for handling untrusted content, instructing the agent to treat instruction-like text in external data as information for the user rather than directives to follow. This is a robust practice for mitigating indirect prompt injection.
[COMMAND_EXECUTION]: While the skill templates mention project commands (e.g., npm run build, npm test), these are provided as examples for project-specific rules files and are not executed by the skill itself.

context-engineering