setup-project
Pass
Audited by Gen Agent Trust Hub on Aug 28, 2026
Risk Level: SAFEREMOTE_CODE_EXECUTIONCOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTIONDYNAMIC_EXECUTION
Full Analysis
- [REMOTE_CODE_EXECUTION]: The skill facilitates the installation of additional capabilities by executing
npx skills add. It references external repositories including Vercel Labs' official agent-skills (a trusted organization), PayloadCMS (a well-known service), and the author's own repository (Firzus/agent-skills). - [COMMAND_EXECUTION]: The skill uses the
gh(GitHub) CLI to automate repository configuration, such as setting squash-merge policies and creating branch protection rulesets. It also usesgitcommands to analyze the repository state. - [INDIRECT_PROMPT_INJECTION]: The skill analyzes external project files (e.g.,
package.json,Cargo.toml,pyproject.toml) to identify the tech stack. Although these files represent an untrusted ingestion point, the skill mitigates risks by requiring explicit user review and confirmation for all proposed changes and command executions. - [DYNAMIC_EXECUTION]: The skill dynamically constructs a JSON payload for GitHub rulesets and writes it to a temporary file (
ruleset.json) before submitting it to the GitHub API. This is a controlled configuration generation process based on predefined templates.
Audit Metadata