The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill is designed to analyze and modify Gradle build files and project configurations, which creates an ingestion surface for potentially untrusted data if a user provides a malicious project.\n
Ingestion points: Project files such as build.gradle.kts and libs.versions.toml are read into the agent context via the Read and Glob tools.\n
Boundary markers: The instructions do not provide explicit delimiters or warnings to the agent to disregard natural language instructions that might be embedded in code comments or string literals within the build files.\n
Capability inventory: The skill is configured with Bash, Write, and Read capabilities, which could be leveraged if an indirect prompt injection attack successfully influences the agent.\n
Sanitization: The skill's validation script (scripts/validate.py) checks for structural integrity and configuration schema but does not sanitize or filter the content of ingested files for prompt injection patterns.

java-gradle