sweeper-plan
Pass
Audited by Gen Agent Trust Hub on Oct 12, 2026
Risk Level: SAFEEXTERNAL_DOWNLOADSCOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
- External Data Retrieval: The skill fetches review records and issue metadata from GitHub repositories (egamma/vscodesweeper-state and microsoft/vscode). These are well-known repositories associated with the VS Code development ecosystem. Retrieving data from these sources is a standard part of the skill's workflow to provide the agent with necessary context.
- Command Execution: The skill utilizes the GitHub CLI (gh) and Git to interact with the repository and verify the environment. Specifically, it uses
gh apiandgh issue viewto gather information andgit rev-parseto manage local configuration. These operations are scoped to the intended functionality of planning issues within a local checkout. - Indirect Prompt Injection Surface: The skill processes external data from GitHub issues and review records, which could potentially contain malicious instructions. However, the skill includes a specific safety rule requiring the agent to treat this content strictly as data, reducing the risk of the agent following embedded commands.
- Configuration Management: The skill modifies the local repository's git exclude file (
info/exclude) to ensure that generated plan files are not accidentally committed or tracked by version control. This is a best-practice measure to keep the working directory clean during the planning phase.
Audit Metadata