The Agent Skills Directory

[PROMPT_INJECTION]: The skill instructs the agent to analyze external data sources including test logs, build outputs, and source code. This creates a surface for indirect prompt injection where malicious instructions embedded in error messages or logs could attempt to manipulate the agent's logic.
Ingestion points: Output from shell commands (e.g., npm test, cargo build) and file content read via the Read tool.
Boundary markers: The skill does not provide specific instructions to the agent to treat external log output as untrusted data or to ignore embedded instructions within that data.
Capability inventory: The agent is authorized to use Bash, Edit, Read, Glob, and Grep tools, providing the capability to modify files and execute commands based on its analysis.
Sanitization: There are no instructions for sanitizing or escaping the content of logs or code before processing it.
[COMMAND_EXECUTION]: The skill directs the agent to utilize the Bash tool to perform essential debugging tasks, such as running reproduction commands (npm test, cargo build), performing binary searches with git bisect, and tracing code execution. These actions are standard for the intended primary purpose of software troubleshooting.

debugging-and-error-recovery