Use this skill to guide an AI assistant working in the Carbon repository on viewing, searching, and editing files with specialized environment tools. It also sets rules for temporary files and when custom programs may be used.
Guidelines and restrictions on shell commands and file manipulation tools for AI assistants.
Run score Not tested
See how to installThis skill sets rules for AI assistants working with files in the Carbon repository. When specialized environment tools are available, it directs the assistant to use view_file for reading, grep_search or find_by_name for searching, and replace_file_content, multi_replace_file_content, or write_to_file for editing. It rules out generic commands such as cat, grep, sed, and patch, as well as custom scripts intended to bypass those restrictions. Temporary programs may be written and run only when no suitable specialized tool exists or when a complex, systematic transformation spans many directories.
Use this skill when asking an agent to investigate or edit files in the Carbon repository and you need to know which tools it should use. It also explains where temporary files and scratch scripts belong and the cleanup required before work ends.
This guidance applies to work in the Carbon repository. Temporary files must be placed in tmp/ at the workspace root and deleted before the work ends.
This overview was written by AI (OpenAI Codex) from the public SKILL.md. Check the original below for the exact behavior.
AI checked the SKILL.md for operations that need care and for setup requirements. The score comes only from actually running the skill (hands-on test).
Not counted in the score. Rated by AI (OpenAI Codex) as of Oct 26, 2026. Treat it as a guide and read the original before installing.
Review the instructions the agent will receive before you install.
AI assistants working on the Carbon repository MUST NOT use legacy or generic UNIX shell search/edit commands when specialized environment tools exist.
cat, less, grep, sed, or other shell utilities for
viewing, searching, or modifying files.patch to write and apply patch files.view_file instead of cat / less.grep_search / find_by_name instead of grep /
find.replace_file_content, multi_replace_file_content,
or write_to_file instead of sed / patch / python edits.You may only write and run temporary programs to modify source code if no semantic tool is applicable or when performing complex, systematic transforms across many codebase directories simultaneously.
Temporary files and scratchpad test scripts created by the assistant during analysis, experiments, or debugging:
tmp/ subdirectory under the workspace root.No reviews yet. Tell others what you thought.
No login needed. Share how it worked for you, ask a question, or note where it did not run.
No comments yet. Be the first to write one.
Use this skill to read GitHub issue details and comments from the command line with `gh`. It covers issues referenced by number or URL, including issues in other repositories, and keeps access read-only so issue state is not changed.
Follow the Carbon project's Bazel workflow through Bazelisk, a version-managing launcher. Learn commands for building targets, running tests and binaries, enabling AddressSanitizer, and refreshing cached state after system dependency changes.
Use this skill to implement builtin functions in the Carbon compiler, from registering a function name and checking its signature to evaluating constant calls and generating LLVM IR for runtime calls. It also covers prelude bindings and focused tests.
Apply Carbon toolchain conventions for formatting Bazel, C++, Carbon, Markdown, and Python files. The skill also covers clear comments and documentation, accurate names, design choices, and concise commit descriptions.
Use this skill to add or revise diagnostics in the Carbon toolchain. It explains how to register message kinds, format typed arguments, attach source labels and context, and verify that diagnostics follow the project’s wording and test coverage rules.
Use `gh`, the GitHub command-line tool, to list, inspect, and search issues, pull requests, repositories, and code. The skill also explains read-only REST and GraphQL queries, pagination, and ways to filter or format returned data.