land-and-deploy
Pass
Audited by Gen Agent Trust Hub on Sep 30, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONDYNAMIC_EXECUTIONCOMMAND_EXECUTIONEXTERNAL_DOWNLOADS
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill processes untrusted input from PR bodies and external website content. Evidence chain: 1) Ingestion points:
gh pr view(PR body) andaside repl(web content). 2) Boundary markers: Untrusted web content is wrapped in markers. 3) Capability inventory: Full shell, file write, and network access. 4) Sanitization: Usesgstack-issue-guardfor PR bodies. - [DYNAMIC_EXECUTION]: The skill executes dynamically generated code in two ways: 1) Using
evalon the output of local binaries (gstack-slug,gstack-diff-scope) to set environment variables. 2) Generating and executing JavaScript for browser automation viaaside repl. - [COMMAND_EXECUTION]: Automates the land-and-deploy workflow by executing multiple shell commands for git operations, GitHub CLI interactions, and platform-specific deployments (e.g., Fly.io, Heroku).
- [EXTERNAL_DOWNLOADS]: Mentions
aside.comas a source for the Aside browser, recommending its use for macOS-based web actions.
Audit Metadata