The Agent Skills Directory

[PROMPT_INJECTION]: The skill possesses a surface for indirect prompt injection. It is designed to validate 'built features' and engage in 'conversational testing', which involves ingesting external, potentially untrusted data.\n
Ingestion points: Untrusted content enters the agent's context through the reading of built feature files and user-provided test inputs.\n
Boundary markers: The skill definition does not provide explicit delimiters or instructions to ignore embedded commands within the data being tested.\n
Capability inventory: The skill has access to high-impact tools including Bash for command execution, Edit and Write for file system modification, and Agent for delegating tasks.\n
Sanitization: No specific sanitization, filtering, or validation logic is defined in this configuration to prevent the execution of instructions embedded in the features under test.

gsd-verify-work