fusion-issue-authoring
Pass
Audited by Gen Agent Trust Hub on Mar 29, 2026
Risk Level: SAFE
Full Analysis
- [COMMAND_EXECUTION]: The skill uses the GitHub CLI (
gh api graphql) for specific issue operations. This usage is restricted to interacting with the official GitHub API and is documented as a fallback mechanism when standard MCP tools are unavailable.- [DATA_EXFILTRATION]: The skill interacts exclusively with the user's configured GitHub environment via official tools. It includes explicit instructions never to request or expose secrets or credentials and uses local temporary files (.tmp/) for drafting content prior to any network transmission.- [PROMPT_INJECTION]: The skill processes user-supplied context to draft issues but mitigates potential indirect injection risks through a mandatory 'Draft first, Review and confirm' workflow. This ensures that the agent never executes a mutation based on potentially untrusted data without explicit user approval of the final payload.- [EXTERNAL_DOWNLOADS]: The skill references the official GitHub MCP server and specific Equinor-owned subordinate skills. All external references target well-known services (GitHub) or the author's own repositories, posing no security risk.
Audit Metadata