The Agent Skills Directory

[SAFE]: The skill implements a professional debugging workflow and encourages the use of established tools (git, npm) for their intended purposes.
[PROMPT_INJECTION]: The skill defines a surface for indirect prompt injection as it requires the agent to read and analyze untrusted external data such as error logs, stack traces, and code diffs. This is an inherent requirement for its primary purpose of debugging. The risk is mitigated by the skill's instruction to use evidence-based reasoning and falsifiable hypotheses.
Ingestion points: SKILL.md Step 2 involves reading error messages, logs, and git diffs.
Boundary markers: No specific boundary markers or delimiters are mandated for log interpolation.
Capability inventory: The skill uses standard development capabilities including shell execution (npm test, git bisect) and network operations (curl, wget) for reproduction loops.
Sanitization: No explicit sanitization of log content is performed before analysis.
[SAFE]: The included utility script scripts/find-polluter.sh is a standard bisection tool used to identify tests that create side effects in a local environment. It does not perform any unauthorized network operations or file modifications outside of the specified test execution flow.

debug-code