The Agent Skills Directory

[COMMAND_EXECUTION]: The skill guides the agent to run local scripts and build wrappers to inspect and verify Kotlin projects.
Evidence: SKILL.md Step 2 calls python3 scripts/inspect-project.py. Step 5 and 7 instruct the use of ./gradlew and ./mvnw wrappers.
[INDIRECT_PROMPT_INJECTION]: The skill's workflow involves reading and processing local project configuration files, which constitutes a surface for indirect prompt injection if the target project contains malicious content designed to influence the agent.
Ingestion points: scripts/inspect-project.py reads and parses build.gradle, pom.xml, and other project files from the provided --root directory.
Boundary markers: No specific delimiters are used to wrap the output of the inspection script when presented to the agent.
Capability inventory: The agent has capabilities to edit source code, modify build configurations, and execute build tasks via shell commands.
Sanitization: The inspect-project.py script returns structured JSON output, which offers basic structural protection but does not sanitize the content of strings parsed from the project files.

kotlin-development