The Agent Skills Directory

[PROMPT_INJECTION] (SAFE): No instructions were found that attempt to bypass safety guidelines or override system constraints. The instructions are focused solely on content style and structure.
[DATA_EXFILTRATION] (SAFE): The skill does not access sensitive file paths (e.g., ~/.ssh, credentials) or perform network operations.
[EXTERNAL_DOWNLOADS] (SAFE): No external dependencies, package installations, or remote script executions are present.
[INDIRECT_PROMPT_INJECTION] (LOW): The skill describes tasks that ingest untrusted data (source code and README files) for documentation generation and review. While this is a theoretical attack surface, the skill lacks dangerous capabilities (like shell execution or network exfiltration) that would make such an injection exploitable.
[COMMAND_EXECUTION] (SAFE): No shell commands, subprocess spawning, or system-level interactions are present. The TypeScript-like snippets provided are high-level abstractions for agent-to-agent communication.

technical-writing