gh-create
Pass
Audited by Gen Agent Trust Hub on Oct 5, 2026
Risk Level: SAFECOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill constructs shell commands (e.g.,
gh repo view,gh issue list,gh issue create) using a repository identifier (OWNER/REPO) provided directly by the user. If the agent does not properly sanitize this input, it could lead to shell command injection. - [INDIRECT_PROMPT_INJECTION]: The skill ingests untrusted user requests to generate GitHub issues, creating a surface for prompt injection.
- Ingestion points: The user's prompt text provided in the
/gh-createcommand. - Boundary markers: The instructions mandate quoting the original prompt verbatim in a collapsed Markdown block (
<details>) and prefixing every line with>to distinguish it from generated content. - Capability inventory: The skill has the ability to execute shell commands via
ghandcurl, read local files, and perform network requests to GitHub's API and upload endpoints. - Sanitization: The skill contains extensive instructions to redact private data (API keys, tokens, local paths, IP addresses) from the output. It also enforces the use of
--body-fileforgh issue createto avoid shell escaping issues with multi-line input.
Audit Metadata