The Agent Skills Directory

[COMMAND_EXECUTION]: The skill instructs the agent to use standard build-tool commands to manage the application lifecycle.
Evidence: Use of ./mvnw quarkus:dev, ./mvnw verify, and quarkus extension add for dependency and environment management (SKILL.md).
[INDIRECT_PROMPT_INJECTION]: The skill's workflow involves reading project configuration and source files, which constitutes a potential attack surface for indirect injection if project files are tampered with.
Ingestion points: Rule 1 in SKILL.md directs the agent to read pom.xml, build.gradle, and existing application source code to determine the environment.
Boundary markers: The skill relies on standard Java and XML/Gradle syntax; no specific instructions are provided to the agent to disregard embedded natural language instructions in these files.
Capability inventory: The agent is authorized to execute shell commands through the build system and modify project files including source code and configuration.
Sanitization: The skill encourages the use of official Quarkus tooling for dependency management rather than manual artifact resolution, which serves as a structural safeguard.
[SAFE]: The skill incorporates several security best practices as core operational rules.
Evidence: Rule 7 in SKILL.md and the 'Secrets' section in references/config-cdi.md explicitly forbid hardcoding credentials and mandate the use of environment variables or secure vault extensions.

quarkus-application-development