The Agent Skills Directory

Indirect Prompt Injection (HIGH): The skill instructs the agent to process untrusted external data while having the capability to execute commands, satisfying the requirements for a high-severity injection surface.
Ingestion points: The workflow requires reading source code, system logs, and data from API/database interactions (Phase 2, Phase 3, and Guidelines).
Boundary markers: There are no instructions or delimiters defined to help the agent distinguish between its own logic and instructions that might be embedded in the code or logs being inspected.
Capability inventory: The agent is explicitly authorized to 'run tests' and eventually 'proceed to implement the proposed solutions', which provides a direct execution and write-access path (Guidelines and Final Output section).
Sanitization: No sanitization or validation of the external content is mentioned before it is processed or used in the execution phase.

performing-systematic-debugging-for-stubborn-problems