The Agent Skills Directory

[PROMPT_INJECTION]: The skill has a surface for indirect prompt injection because it interacts with untrusted external data sources such as API endpoints and specifications. * Ingestion points: Data enters the context from external servers via tools like curl, hurl, k6, and nuclei during testing or reconnaissance phases (e.g., OpenAPI specs, GraphQL introspection, or API response bodies). * Boundary markers: The instructions do not explicitly mandate the use of delimiters or specific warnings to ignore instructions embedded within these external data responses. * Capability inventory: The skill is granted extensive capabilities including Bash, Write, Read, Glob, and Grep. * Sanitization: External data is typically piped to jq or written to local report files without robust sanitization against potential instruction-carrying payloads.
[EXTERNAL_DOWNLOADS]: The skill toolchain references several third-party security tools and frameworks. * Evidence: Mentions of @usebruno/cli, hurl, k6, httpie, and nuclei. * Status: These references target well-known open-source projects and official registries (NPM, Cargo, Homebrew). These downloads are standard for the skill's purpose and originate from trusted service providers.

api-testing