The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill is designed to ingest and analyze untrusted source code from a project repository. A malicious actor could embed instructions in code comments or test files to influence the agent's behavior during analysis.
Ingestion points: Project source files (TypeScript, JavaScript, Python, Java, Kotlin, Go, C#) within the repository being analyzed.
Boundary markers: No specific boundary markers or "ignore instructions" delimiters are implemented for the data ingestion process.
Capability inventory: The skill utilizes various shell commands including test runners (vitest, pytest), linters (eslint, ruff, golangci-lint), and custom analysis scripts (python3 tools/cc-scan.py).
Sanitization: No sanitization or filtering of the external source code is performed before processing or executing analysis tools.
[COMMAND_EXECUTION]: The skill provides detailed instructions for the agent to execute shell commands to perform code quality scans, formatting, and unit testing. These commands are intended to be run on the local codebase.
[EXTERNAL_DOWNLOADS]: The instructions suggest using npx to run tools like Prettier and ESLint, and mentions Python packages like Ruff and Pytest. These tools are fetched from well-known public registries (NPM and PyPI). These references are documented neutrally as they target standard services for the skill's primary development purpose.

clean-code