egraph-rs-dev-process

name	egraph-rs-dev-process
description	Guidelines for general development process, coding style, writing comments, and updating skills after tasks in the egraph-rs project.

Development Process Guidelines for egraph-rs

This skill details the general development rules, comment conventions, task workflow, and self-updating skill process for the egraph-rs project.

Code Comment Rules

• Language: All comments in code files (Rust, Python, JS, TS) must be written in English.
• Content: Write comments that explain the WHY, not the HOW. Explain design decisions, algorithms, and logical justifications.
• Cleanliness: Avoid unnecessary commented-out code or work-in-progress notes.

Task Execution Workflow

1. Before Starting a Task

• Understand task instructions and constraints.
• Refer to egraph-rs-system-patterns skill to understand the project architecture.
• Examine related code files to understand existing implementation patterns.
• Create an internal task checklist (e.g. task.md) to plan and track your work.

2. During Implementation

• Ensure all comments adhere to the "Code Comment Rules" above.
• Follow formatting and style guidelines.
• Regularly verify changes.

3. Before Task Completion (Final Verification Process)

• Formatting and Linting:
  • Run cargo fmt --all and ensure no formatting changes remain.
  • Run cargo clippy --workspace --all-targets --all-features -- -D warnings and ensure there are no warnings.
• Run Tests:
  • Verify that the changes do not break existing code by running workspace tests.
  • See egraph-rs-verification skill for specific test commands.
• Prepare Commit:
  • Draft a commit message following Conventional Commits format (see egraph-rs-commit skill).
• Ask for User Confirmation:
  • Present a summary of all changes made.
  • Include the proposed commit message.
  • Wait for explicit user approval before completing the task. This confirmation step must never be skipped under any circumstances.

Post-Task Reflection: Updating Skills

Upon completing any task, the agent MUST reflect on the work performed.

• Analyze if new implementation patterns, codebase quirks, build command requirements, or general instructions were learned during the task.
• If new patterns or insights emerged, immediately update the relevant skill files (e.g., egraph-rs-dev-process, egraph-rs-verification, egraph-rs-system-patterns) to keep the guidelines up-to-date.
• This self-updating loop replaces the legacy "Memory Bank" mechanism and ensures future agents benefit from the context established in this task.