hotfix
Pass
Audited by Gen Agent Trust Hub on Oct 15, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill processes external data from bug descriptions and repository-local documentation while possessing the capability to execute shell commands and modify files.
- Ingestion points: Reads bug IDs, descriptions, and project-specific instruction files (
AGENTS.md,CLAUDE.md) to inform the hotfix process. - Boundary markers: While it does not utilize explicit delimiters for external content, the workflow is partitioned into distinct phases with mandatory
AskUserQuestioncheckpoints that provide manual oversight. - Capability inventory: Employs
Bash,Write,Edit, andTasktools, which allow for repository modification and subprocess execution. - Sanitization: User-provided identifiers are interpolated into shell commands (e.g., git branch names), but these are mitigated by displaying the final command to the user for approval before execution.
Audit Metadata