The Agent Skills Directory

[EXTERNAL_DOWNLOADS]: The skill downloads and installs standard open-source developer utilities from official registries and repositories.
Fetches madge via npm/bunx for JavaScript/TypeScript analysis.
Installs pycycle and import-linter via pipx for Python analysis.
Installs goda from its official GitHub repository via go install.
Installs cargo-modules from crates.io via cargo install.
[COMMAND_EXECUTION]: Executes shell commands to perform project builds and run verification suites.
Runs go build and cargo build to confirm project integrity.
Executes project-defined scripts such as bun run check and pytest to verify refactorings.
Uses CLI tools like madge and pycycle to generate dependency reports in JSON or text formats.
[PROMPT_INJECTION]: The skill possesses an indirect prompt injection surface as it processes untrusted source code to guide its refactoring logic and command execution.
Ingestion points: Reads local source code files (TS, JS, PY, GO, RS) during dependency analysis in SKILL.md.
Boundary markers: No specific delimiters or instructions to ignore embedded content are present for the analyzed files.
Capability inventory: The skill can create and modify files, and execute shell commands for builds and testing as seen in the Apply and Verify sections.
Sanitization: Does not explicitly sanitize file paths or content extracted from the dependency analysis before using them in file operations or shell commands.

cleanup-cycles