The Agent Skills Directory

[COMMAND_EXECUTION]: The skill documentation and referenced guidelines specify the use of Docker (via testcontainers) and local command runners (just, make) to execute integration and E2E tests. This is a primary function of the skill, requiring environment permissions to manage containers and run shell commands.
[INDIRECT_PROMPT_INJECTION]: The testing framework supports a template-based test data pattern that parses files to generate test inputs, creating a surface for potential instruction injection if template sources are untrusted.
Ingestion points: Template files are read from the templates/ directory using template.ParseFiles as documented in references/fluent-testing-guideline.md.
Boundary markers: No explicit delimiters or instructions to ignore embedded commands are described in the template processing examples.
Capability inventory: The skill possesses extensive capabilities including database execution (db.Exec), network requests (http.Client), SQS messaging (sqsClient), and container management (testcontainers.GenericContainer).
Sanitization: The provided implementation for sendMessageFromTemplate does not show explicit sanitization or validation of the template content before execution.
[CREDENTIALS_UNSAFE]: Code examples in SKILL.md include hardcoded credentials for local testing environments, such as POSTGRES_USER: "test" and POSTGRES_PASSWORD: "test". While appropriate for local development, these represent a pattern of handling sensitive information within the skill instructions.

go-integration-tests