The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill is designed to read and analyze project-specific files which serve as untrusted external inputs that could contain malicious instructions.
Ingestion points: The agent is instructed to read SQL migration scripts from src/main/resources/db/migration and configuration files such as application.yml or pom.xml.
Boundary markers: The instructions do not define specific delimiters or "ignore" markers for content within the ingested SQL or configuration files.
Capability inventory: The skill has the capability to modify project files and execute shell commands via the Maven wrapper (./mvnw compile, ./mvnw clean verify).
Sanitization: There is no explicit logic to sanitize or filter out instructions hidden in SQL comments or YAML metadata.
[COMMAND_EXECUTION]: The skill workflow involves executing shell commands to ensure project integrity and verify changes.
Evidence: SKILL.md contains mandatory instructions to run ./mvnw compile before applying changes and ./mvnw clean verify after improvements. While these are standard development practices, they represent a capability that could be targeted if the agent is influenced by malicious input.

513-frameworks-micronaut-db-migrations-flyway