The Agent Skills Directory

[SAFE]: No malicious patterns or security vulnerabilities were detected. The skill guides the agent through a logical process of reproducing bugs via local tests, identifying root causes within the project's source files, and verifying fixes using standard build tools like npm.
[INDIRECT_PROMPT_INJECTION]: While the skill processes external bug reports, which are untrusted data, the workflow requires the agent to verify the symptom with a localized test case and map it to specific source code lines (root cause). These steps provide a strong logical buffer against adversarial inputs attempting to influence the agent's behavior beyond its intended scope.

fix-bug