comet-github-idea-to-issue
Pass
Audited by Gen Agent Trust Hub on Aug 14, 2026
Risk Level: SAFECOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
- [PROMPT_INJECTION]: The skill ingests untrusted data from the local environment and external GitHub search results to generate issue drafts, creating a surface for indirect prompt injection.
- Ingestion points: Reads local code, tests, documentation, and external GitHub issues/PRs as specified in
SKILL.md. - Boundary markers: The skill instructions do not provide strict delimiters or sanitization rules for the external content it processes, although it suggests labeling unverified content as hypotheses.
- Capability inventory: Reads local files and performs write operations via the GitHub CLI.
- Sanitization: Requires explicit user confirmation before executing the
gh issue createcommand, which acts as a human-in-the-loop control. - [COMMAND_EXECUTION]: The skill uses the GitHub CLI (
gh) to create new issues. - Evidence: Mentions invoking
gh issue createinSKILL.md.
Audit Metadata