The Agent Skills Directory

[COMMAND_EXECUTION]: The skill executes python -m skore_skills status at the beginning of its procedure to retrieve the current state of the workspace. This uses a project-specific tool associated with the skill's environment.
[EXTERNAL_DOWNLOADS]: The skill fetches primary pages from scientific and technical sources as part of its research functionality. This is the primary intended use of the skill.
[INDIRECT_PROMPT_INJECTION]: The skill processes untrusted data from external websites which could theoretically contain instructions designed to influence the agent.
Ingestion points: Content retrieved from external search results (primary pages).
Boundary markers: No specific delimiters are defined for the fetched content.
Capability inventory: The agent has the ability to write to files (journal/ideas/, scratch/research/) and execute workspace status commands.
Sanitization: The skill instructions require rewriting the search query to remove specific dataset names or loader identifiers before performing the search, which limits the exposure of internal project details to external services.

search-ml-literature