The Agent Skills Directory

[SAFE]: The skill implements a structured debugging methodology using four distinct phases: Root Cause Investigation, Pattern Analysis, Hypothesis and Testing, and Implementation. These instructions guide the agent toward systematic problem-solving.
[SAFE]: The utility script find-polluter.sh is a standard diagnostic tool for identifying tests that cause state pollution. It uses common shell commands (find, npm test, ls) to isolate problematic test files.
[SAFE]: Diagnostic instrumentation examples provided in SKILL.md and root-cause-tracing.md follow security best practices. For instance, the script example for checking secrets uses shell parameter expansion (${IDENTITY:+SET}) to verify the existence of a secret without actually logging or exposing its value.
[SAFE]: The skill does not contain any patterns associated with prompt injection, data exfiltration, obfuscation, or persistence. The 'Iron Law' and other strict instructions are functional constraints designed to ensure debugging quality and do not attempt to bypass system safety filters.

systematic-debugging