The Agent Skills Directory

[COMMAND_EXECUTION]: The skill instructs the agent to execute shell commands using uv run researchkit. This is the primary mechanism for running the research pipeline and handles dependency management.
[INDIRECT_PROMPT_INJECTION]: The skill operates by researching external web content and processing 'materials', which are then synthesized into reports. This creates a surface where data from untrusted web sources could contain instructions designed to influence the agent's summary or behavior.
Ingestion points: External web search results and documents provided via the --materials or --ingest flags (found in SKILL.md Workflow section).
Boundary markers: None explicitly defined in the instructions to separate untrusted web content from agent instructions.
Capability inventory: The skill has the capability to execute commands (uv run) and write files to the local projects/ directory (found in SKILL.md Workflow section).
Sanitization: No specific sanitization or filtering of external web content is described in the skill configuration.

explore