The Agent Skills Directory

[COMMAND_EXECUTION]: The skill instructs the agent on how to use standard CLI tools for package management and testing, such as npm install, npx jest, and npx vitest. These commands are standard for the described development tasks.
[EXTERNAL_DOWNLOADS]: The instructions reference standard, well-known libraries from the official npm registry and common GitHub Actions like codecov/codecov-action. These are documented neutrally and follow industry best practices for CI/CD and dependency management.
[DATA_EXFILTRATION]: No exfiltration patterns were detected. References to environment variables and secrets (e.g., CODECOV_TOKEN) occur within standard CI/CD configuration snippets where such usage is expected and follows secure patterns.
[INDIRECT_PROMPT_INJECTION]: The skill operates on project source code and configuration files. While this creates a surface for indirect prompt injection if an agent were to process a malicious repository, the skill's instructions focus strictly on testing logic and do not implement any patterns that would facilitate or encourage the execution of instructions found within data.
[OBFUSCATION]: A thorough review of all 25 files, including the extensive evaluation datasets, revealed no evidence of obfuscated code, hidden characters, or deceptive content.

js-unit-tests