The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill is designed to ingest security requirements (NFRs) from an external agent (qa-nfr-analyst) and transform them into executable TypeScript or Python code, which presents a risk of processing malicious instructions embedded in the data.
Ingestion points: The workflow in SKILL.md describes reading NFR analysis to map test scenarios.
Boundary markers: There are no defined delimiters or instructions to ignore embedded commands within the ingested requirement data.
Capability inventory: The skill has the capability to generate and write executable scripts (Python, TypeScript) and configuration files (YAML) to the tests/security directory.
Sanitization: The provided instructions do not describe sanitization or validation of external input before it is used for code generation.
[COMMAND_EXECUTION]: The skill provides documentation and examples for executing commands via shell or CLI, including Docker commands for running OWASP ZAP (docker run -t owasp/zap2docker-stable) and system calls within generated test scripts (exec('npm audit --audit-level=high')). While the skill specifies that the user executes these, the provision of exact command strings represents a potential execution vector.
[EXTERNAL_DOWNLOADS]: The documentation references external resources including the official owasp/zap2docker-stable Docker image. It also specifies dependencies on several common testing frameworks and libraries such as Playwright, Supertest, httpx, requests, and pytest for the resulting security suite.
[CREDENTIALS_UNSAFE]: Reference files include examples with placeholder API keys (your-api-key) and dummy credentials (TestPass123!) in references/zap-config.md. The skill correctly advises using environment variables and .env files for real secrets in its best practices documentation, reducing the risk of accidental exposure of real credentials.

qa-security-test-writer