The Agent Skills Directory

[COMMAND_EXECUTION]: The skill utilizes shell commands to verify the prototype's functionality.
Evidence: The procedure explicitly instructs the agent to launch a local development server using npm run dev or execute test scripts.
[INDIRECT_PROMPT_INJECTION]: The skill ingests external documentation and models, creating a potential vector for indirect instructions.
Ingestion points: The skill processes hypothesisToValidate, domainModel, and researchReport in SKILL.md.
Boundary markers: Absent; there are no specific instructions to delimit or ignore instructions embedded within the research or model data.
Capability inventory: The skill allows file writing (creating prototype-report.md and code spikes) and shell command execution (npm run dev).
Sanitization: Absent; no validation or escaping mechanisms are mentioned for the ingested research data.
[DYNAMIC_EXECUTION]: The skill involves the generation and execution of temporary code spikes.
Evidence: The 'Procedure' and 'Outputs' sections describe the creation of a 'lightweight disposable code spike' and its subsequent execution for validation purposes.

prototype