The Agent Skills Directory

[EXTERNAL_DOWNLOADS]: The skill uses npx to download and execute the @agent-analytics/cli package. It also includes instructions that download an initialization script via bash <(curl...). Per the [TRUST-SCOPE-RULE], these operations target a well-known vendor ('Agent Analytics') and are documented as safe.
[COMMAND_EXECUTION]: The skill executes shell scripts and CLI commands (npx, bash, perl) to manage experiment data, initialize directories, and process logs. These commands are local to the agent's environment and perform intended administrative tasks related to the growth loop.
[SAFE]: No malicious patterns such as prompt injection, credential theft, or unauthorized data exfiltration were detected. The skill follows best practices by requesting human approval before implementing product changes.

agent-analytics-autoresearch