big-code-analysis

End-to-end workflow for adding a new tree-sitter language to big-code-analysis. Wires up the grammar crate, regenerates the language enum, implements Checker/Getter/Alterator/metrics, adds per-metric tests, updates docs, then runs simplify-rust, rust-optimize, and audit-tests. Does NOT commit — leaves the working tree dirty for final user review.

Audit a Rust crate in the big-code-analysis workspace for logic errors, complexity, bugs, security issues, and code smells. Operates at the crate level via `cargo -p`. Use when asked to audit or review a crate.

Audit a single Rust source file in the big-code-analysis workspace for logic errors, complexity, bugs, security issues, and code smells. Use when asked to audit or review a specific file.

Audit naming quality in a crate or directory for misleading, inconsistent, or unclear names. Use when asked to audit or review naming.

Audit test suites for tests that pass trivially, mask bugs, or assert the wrong thing. Finds tests designed to pass rather than designed to catch regressions.

Fix multiple GitHub issues on an integration branch. Issues touching different crates run in parallel worktrees; issues sharing a crate or affecting cross-language code run sequentially. Each goes through fix, simplify, review, and remediation before merging. Use when asked to fix several issues at once.

Remove dead code, unused imports, and unreachable paths from a single Rust crate. Creates an integration branch and applies removals in parallel worktrees.

Author or revise a human-consumed prose document (README, STABILITY, docs under docs/ and the mdBook under big-code-analysis-book/src/): clarify audience and purpose, outline, draft against docs/conventions/documentation.md, then revise in multiple focused passes (structure, completeness, line edit, proofread) that hunt the three doc failure modes (missing external links, insufficient background, hollow sounds-smart sentences), get an independent fresh-context review, and verify markdown lint. Includes a review mode that critiques an existing doc without editing it. Use when asked to write, revise, tighten, or de-slop a documentation or Markdown prose file. Not for code, diffs, doc comments, changelog entries, or lessons-learned (those have their own workflows).

Complete workflow for fixing GitHub issues including investigation, implementation, review, testing, and documentation. Use when asked to fix a GitHub issue.

Safe code improvement workflow for a single crate. Creates an integration branch, analyzes code quality, and applies fixes in parallel worktrees.

Read a GitHub issue, develop a detailed resolution plan using sequential thinking, rate difficulty/complexity/priority, and label low-priority issues. Use when asked to plan or assess an issue.

Fetch open GitHub issues and generate a triage report with quick wins and recommended groupings. Use when asked to triage issues.

Audit code changes for correctness, performance, security, and quality issues. Use when asked to review changes, diffs, or pull requests.

Analyzes Rust code to reduce verbosity and line count while preserving behavior. Applies modern idiomatic patterns, newer syntactic features (up to Rust 1.94), and better type-system modeling including newtypes, enums, and trait design. Use when asked to simplify, shorten, optimize, clean up, or modernize Rust code.

Scan of the core library, CLI, and web crates for logic errors, security issues, and metrics calculation bugs across all language modules. Runs 6 parallel sub-agents, one per scope partition, and applies a 50-question checklist distilled from the highest-recurrence lessons in docs/development/lessons_learned.md. Files GitHub issues with fix plans. READ-ONLY — no commits, no file modifications.

Review changed Rust code for reuse, quality, and efficiency, then fix any issues found.

Review project activity and draft entries for lessons_learned.md. Use when asked to update or review lessons learned.