The Agent Skills Directory

[Command Execution]: The skill utilizes shell commands for build verification. It proactively addresses command injection risks by instructing the agent to validate user-provided identifiers against a strict regular expression and ensuring path resolution remains within the intended directory.
[Data Sanitization]: The instructions mandate a redaction step before recording bug details in the repository state file. This process is designed to remove sensitive information like tokens, PII, and internal hostnames, preventing data leaks in persistent logs.
[Investigation Integrity]: By requiring a trace of the failure to specific lines of evidence before proposing a fix, the skill ensures that modifications are grounded in verifiable code, reducing the risk of the agent being misled by indirect injection in bug reports.

debug-extension