Clean up AI-generated writing by removing slop (generic tone, filler, inflated claims, vague attributions, list spam) while keeping meaning the same. Use when asked to "humanize" text, remove AI-sounding writing, or align prose with local/author voice.
Write and edit in a clear, direct style that avoids AI-sounding artifacts. Use whenever drafting or rewriting prose (docs, READMEs, PR descriptions, issues, blog posts, emails, UI copy) and when aligning to the author's preferred voice.
Clean up AI-generated code changes by reviewing diffs against main and removing stylistic or structural artifacts (over-commenting, unnecessary defensive checks, try/catch noise, any-casts, inconsistent style). Use when asked to “remove AI code slop,” “humanize” changes, or align a branch’s edits with existing codebase conventions.
Open pull requests for the current git repo: verify you are not on main/prod, compare your branch to main, commit any pending work, and create a PR (prefer gh). Use when asked to open or create a PR, draft a PR description, or automate the PR workflow for a local git branch.
Refactor AGENTS.md instructions to follow progressive disclosure and AGENTS.md best practices. Use when asked to audit, reorganize, or split AGENTS.md into a minimal root file plus linked docs; includes contradiction checks, essential extraction, grouping, and deletion flags.
Plan commits from the current working tree. Use when asked how to split changes into commits, what to stage together, how to handle untracked files, or to propose commit messages and an execution plan (stage/commit) based on `git status`/diffs.
General coding principles for writing clean, maintainable code. Use for development work, code reviews, refactoring, dependency changes, and design discussions about code structure, performance, and reliability.
Safe-by-default git workflow and Conventional Commits. Use for git operations (diffs, staging, commits, branches, merges, conflict resolution, rebase/cherry-pick), PR review, and releases (tagging). Prevent destructive operations without consent.