The Agent Skills Directory

[COMMAND_EXECUTION]: The skill executes shell commands via ./gradlew to run specialized codebase tests, compile modules, and retrieve file diffs using git commands.
[PRIVILEGE_ESCALATION]: To enable the auto-apply feature for code modifications, the skill sets the property -Pkotlin.test.instrumentation.disable.inputs.check=true, which explicitly disables the security manager that would normally prevent the test JVM from writing to the local source files.
[INDIRECT_PROMPT_INJECTION]: The skill reads and processes content from source code files, git diff outputs, and search results from an external repository path provided by the user. This data ingestion creates an attack surface where malicious instructions embedded in the codebase could attempt to influence the agent's behavior.
Ingestion points: git diff output, mcp__idea__read_file results, and Grep results from the external IntelliJ path.
Boundary markers: None present; the skill treats modified code blocks and search results as trusted data for analysis.
Capability inventory: Executes shell commands (Gradle), reads/writes files, and applies patches to the codebase.
Sanitization: No explicit sanitization or escaping of external content is mentioned before it is processed by the agent's logic.
[COMMAND_EXECUTION]: The skill accepts a user-supplied directory path via the --intellij flag, which is used as a target for the Grep command. This allows the execution of search operations against arbitrary file system paths provided by the user.

analysis-api-mark-internal-apis