The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill analyzes repository files such as .env, README.md, and GitHub workflow files to determine the necessary configuration steps. This process involves ingesting untrusted local data, which represents a potential attack surface where malicious content in those files could influence the generated script's logic.
Ingestion points: The agent is instructed to read .env, .env.example, README, docker-compose.yaml, and .github/workflows/* (SKILL.md).
Boundary markers: No explicit delimiters or warnings are used when processing content from these files.
Capability inventory: The resulting scripts can write to local files, set GitHub secrets via the gh CLI, and open browser URLs (template.sh).
Sanitization: The generated scripts use ask_secret to hide sensitive user input, but the extraction of keys from the repository relies on the agent's parsing logic.
[EXTERNAL_DOWNLOADS]: The template script includes an example that references Stripe's official dashboard for API key management. This is a reference to a well-known service intended for infrastructure setup.
[COMMAND_EXECUTION]: The generated scripts utilize standard shell utilities (grep, mktemp) and the GitHub CLI (gh) to automate configuration tasks. These operations are performed at the user's discretion when running the generated script.

wizard