The Agent Skills Directory

[SAFE]: The skill is a set of instructional guidelines designed to improve debugging outcomes by enforcing a systematic approach. It does not include any scripts, binaries, or automated network operations.
[COMMAND_EXECUTION]: The instructions direct the agent to use standard, read-only development commands like grep and git log to search for error patterns and review recent code changes. These are well-defined within the scope of a technical debugging task.
[PROMPT_INJECTION]: The skill uses authoritative terminology such as the "Iron Law" and "MANDATORY" steps. These are functional constraints intended to discipline the AI's technical reasoning (preventing "guess-and-check" fixes) and do not represent attempts to bypass system safety filters or extract system prompts.
[DATA_EXFILTRATION]: No sensitive file access or network exfiltration patterns were detected. All data analysis (logs, stack traces) remains within the agent's local working context.

Systematic Debugging