The Agent Skills Directory

[SAFE]: The skill is a documentation resource for building and managing PostHog Signals scouts. It contains no executable code or scripts that perform dangerous operations. All referenced tools and domains (e.g., posthog.com, github.com) are legitimate and consistent with the skill's stated purpose and author (posthog).
[INDIRECT_PROMPT_INJECTION]: The skill provides explicit safety guidance (in references/scout-patterns.md) regarding indirect prompt injection. It instructs authors to treat ingested content from external sources (like warehouse tables, Slack, or GitHub) as untrusted data and never as instructions, significantly reducing the risk of injection attacks in the scouts it helps author.
[DATA_EXPOSURE]: The documentation includes guidelines on sanitizing PII (Personally Identifiable Information) when processing open-text data, demonstrating a security-first approach to data handling.
[COMMAND_EXECUTION]: While the skill describes tools for command execution (like hogli or gh), these are documented in the context of developer workflows and testing, and no malicious commands are present in the instruction set.

authoring-scouts