The Agent Skills Directory

[COMMAND_EXECUTION]: The skill includes a local utility script 'scripts/bootstrap-experience.ts' used for initializing the project's experience tracking directory. This script contains safety logic to handle file system operations securely.
[EXTERNAL_DOWNLOADS]: The research agent utilizes platform-provided web search tools to gather information, which is a core part of its documented functionality.
[PROMPT_INJECTION]: The skill is designed to handle untrusted data from the web by implementing a sequential pipeline where a separate 'critic' agent evaluates the draft against a strict rubric, specifically checking for originality and evidence quality.

write-longform