The Agent Skills Directory

[Identity and Permission Verification]: The skill includes a dedicated step to verify the active GitHub account and its permissions (push or triage) before attempting any operations. This prevents actions from being taken under the wrong identity or failing due to insufficient permissions.- [Dry-Run and Approval Workflow]: A core security feature of this skill is its mandatory dry-run phase. It generates a detailed plan of proposed assignments and requires explicit user approval before making any changes to the repository state.- [Secure Configuration Management]: Reviewer configuration is handled through a structured reviewers.json file governed by a JSON schema. The skill implements strict validation to ensure the configuration is well-formed and consistent, preventing errors or unexpected behavior from malformed settings.- [Input Sanitization and Logic]: The skill uses specific logic for area detection based on file paths and conventional commit scopes, with a clear hierarchy that prioritizes authoritative data (paths) over user-supplied text (titles). It also includes deduplication and author-exclusion logic to prevent redundant or invalid requests.- [Least Privilege and Transparency]: The skill explicitly states it does not perform approvals, merges, or comments, limiting its scope to reviewer requests only. It avoids hardcoding sensitive data and relies on the user's existing authenticated environment.

assign-prs