The Agent Skills Directory

[COMMAND_EXECUTION]: The skill uses dynamic context injection via the !command`` syntax in SKILL.md to execute shell commands like pwd, find, and grep at load time. These commands are used for project introspection to automatically detect the language, framework, and existing test configuration.
[EXTERNAL_DOWNLOADS]: The skill automates the installation of standard testing dependencies including supertest, testcontainers, and httpx using package managers like bun and uv. These are well-known libraries from official registries.
[COMMAND_EXECUTION]: It provides instructions to modify and create GitHub Actions workflow files (.github/workflows/test.yml) to integrate testing jobs with service containers (PostgreSQL, Redis).
[PROMPT_INJECTION]: The skill possesses a surface for indirect prompt injection as it ingests and acts upon data from untrusted project manifest files.
Ingestion points: Project configuration files such as package.json, pyproject.toml, and .project-standards.yaml are read to analyze the testing state.
Boundary markers: No delimiters or warnings are provided to prevent the agent from obeying instructions potentially embedded in these ingested files.
Capability inventory: The skill is capable of modifying project files and executing shell commands (Bash), which could be misdirected by a malicious project manifest.
Sanitization: There is no evidence of sanitization or content validation before these files influence the agent's logic.

configure-integration-tests