The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill is designed to ingest and analyze untrusted Java source code, creating an inherent attack surface for indirect prompt injection.
Ingestion points: The agent analyzes Java source files (*.java) provided by the user or located in the workspace (referenced in SKILL.md analysis commands).
Boundary markers: There are no explicit instructions to use delimiters or ignore instructions embedded in code comments during the review process.
Capability inventory: The skill leverages shell commands (grep) to perform static analysis of the codebase.
Sanitization: The skill does not define specific sanitization or filtering logic for the code it processes.
[COMMAND_EXECUTION]: The skill provides the agent with specific shell commands to facilitate the analysis of Java code in the local workspace.
Evidence: SKILL.md includes commands such as grep -rn "synchronized" --include="*.java", grep -rn "@Async" --include="*.java", and grep -rn "volatile" --include="*.java" for static analysis purposes.

concurrency-review