The Agent Skills Directory

[COMMAND_EXECUTION]: The skill defines a series of bash functions that execute standard command-line tools such as npm, npx, git, and python3. These are used for tasks like linting, running tests, checking build logs, and auditing the local repository for potential secrets.
[EXTERNAL_DOWNLOADS]: It utilizes curl to perform network requests against a user-specified environment variable ($PROD_URL). These requests are used to verify HTTP status codes, inspect security headers (e.g., HttpOnly and Secure flags on cookies), and retrieve HTML content for SEO validation.
[PROMPT_INJECTION]: The skill exhibits an indirect prompt injection surface (Category 8) by ingesting and processing content from an external production URL. While these operations are standard for a QA tool, the data retrieved could theoretically contain malicious instructions designed to influence the agent's interpretation of the results.
Ingestion points: Content is fetched from the user-provided $PROD_URL within functions like qa:seo, qa:api, and qa:smoke (found in SKILL.md).
Boundary markers: The skill does not employ explicit delimiters or instructions to ignore embedded commands within the fetched HTML or JSON data.
Capability inventory: The skill utilizes shell execution, file system access for local scanning, and network operations via curl (documented in SKILL.md).
Sanitization: Data retrieved from the external URL is processed using utilities like grep and python3 without specific sanitization or escaping of the raw content.

vibecode-production-qa-validator