init-bug-tracker
Pass
Audited by Gen Agent Trust Hub on Sep 12, 2026
Risk Level: SAFECOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill uses the GitHub CLI (gh) and Git for core operations, such as creating labels, cloning repositories, and pushing template updates. These commands are clearly defined and perform the intended developer-oriented tasks.
- [INDIRECT_PROMPT_INJECTION]: The skill analyzes local repository content to generate suggestions, which establishes a surface for indirect instructions to influence the agent.
- Ingestion points: Step 2 reads the local README.md, directory layout, and CONTEXT.md to derive label slugs and descriptions.
- Boundary markers: There are no explicit delimiters or specific 'ignore instructions' markers applied to the text read from the source files.
- Capability inventory: The skill has access to the shell (gh and git tools), the ability to write to agent configuration files (CLAUDE.md and AGENTS.md), and the ability to perform network operations via Git push.
- Sanitization: The skill includes a significant human-in-the-loop control, requiring the maintainer to confirm, edit, or replace the proposed labels before they are created or updated in the target repository.
Audit Metadata