Use OpenAI Codex CLI specifically to get a second opinion. Trigger only when the user explicitly asks for a Codex review.
Best practices for maintaining a forked git repository in sync with the upstream original, and for creating branches to submit pull requests. Use when the user asks about: keeping a fork up to date, syncing with upstream, contributing to an open source project via fork, creating PR branches from a fork, fork workflow, "my fork is behind upstream", "how do I contribute to a repo I forked", any question about fork/upstream/PR git workflow, or how to run all local feature branches together while some are still pending merge upstream.
Bootstrap a new Python project with the Astral toolchain (UV, Ruff, ty), flake8-stepdown, and pre-commit hooks. Opinionated.
Run parallel code reviews using Claude, OpenCode, and Codex, then merge findings into a single plan.
Run OpenCode to review uncommitted changes, then build an action plan from its findings.
Own this plugin?
Verify ownership to unlock analytics, metadata editing, and a verified badge. GitHub access is read-only (username + org membership).
Sign in to claimOwn this plugin?
Verify ownership to unlock analytics, metadata editing, and a verified badge. GitHub access is read-only (username + org membership).
Sign in to claimBased on adoption, maintenance, documentation, and repository signals. Not a security audit or endorsement.
A collection of Claude Code skills.
/plugin marketplace add didmar/skills
Skills are then available with the didmar: prefix, e.g. /didmar:forked-repo-git-practices.
forked-repo-git-practicesBest practices for maintaining a forked git repository in sync with its upstream original, and for contributing back via pull requests.
Covers: initial setup, syncing your fork, creating PR branches, keeping PRs updated during review, post-merge cleanup, local integration branches, and common pitfalls.
init-pythonBootstrap a new Python project with the Astral toolchain (UV, Ruff, ty) and pre-commit hooks. Opinionated.
Supports both app and library layouts. Sets up linting, testing (pytest), pre-commit hooks, and optionally: Dockerfile + docker compose, GitHub Actions CI, VS Code settings, pytest-cov, PyPI packaging (library only), and CONTRIBUTING/AGENTS docs.
Usage: /didmar:init-python or /didmar:init-python lib for library layout.
setup-deploy-keyGenerate an SSH deploy key for the current project and configure git to use it. Idempotent.
Creates an ed25519 SSH key pair, adds an SSH config entry with a project-specific host alias, updates the git remote URL, and displays the public key with instructions for adding it to GitHub.
toulminifyExtract and structure arguments from text or URLs using Toulmin's model of argumentation.
Accepts a URL, file path, or raw text as input and identifies every distinct argument, breaking each down into claim, grounds, warrant, backing, qualifier, and rebuttal.
Usage: /didmar:toulminify <url|filepath|text> or /didmar:toulminify --json <url|filepath|text> for JSON output.
recover-sessionsRecover Claude Code sessions after moving a project to a new directory.
When a project is moved, Claude Code sessions become invisible because they are stored under a path-derived directory in ~/.claude/projects/. This skill copies sessions from the old path's storage into the current path's storage.
Usage: /didmar:recover-sessions ~/old-project-path or /didmar:recover-sessions --dry-run ~/old-project-path to preview.
multi-reviewRun three independent code reviews in parallel — using Claude, OpenCode, and Codex — then merge the results into a single deduplicated action plan.
Findings flagged by multiple reviewers get higher confidence. Results are grouped by severity and presented in plan mode for approval before any edits.
Usage: /didmar:multi-review
codex-reviewRun OpenAI Codex CLI to review all uncommitted changes, then enter plan mode so Claude can act on the findings.
Uses Codex's built-in review --uncommitted command with the Landlock sandbox backend to avoid bwrap permission issues in VMs and containers.
Usage: /didmar:codex-review
opencode-reviewRun OpenCode CLI to review all uncommitted changes, then enter plan mode so Claude can act on the findings.
Uses opencode run with a review prompt since OpenCode doesn't have a built-in review command. Model is configurable via opencode.json or -m flag.
Usage: /didmar:opencode-review
npx claudepluginhub didmar/skills --plugin didmarGit and PR workflows with code review, onboarding, and PR enhancement
Coding workflow skills: intelligent commits, push & PR, branch cleanup, CLAUDE.md maintenance, README generation, changelog creation, README updating, and GitHub Actions setup
Development environment setup: git worktrees, terminal optimization
General purpose Claude Code toolkit with research agents, skills, and introspection commands
Iterative Ralph loops with Codex CLI review gates at completion. Claude works until done, then Codex reviews before allowing exit.
Ultra-compressed communication mode. Cuts ~75% of tokens while keeping full technical accuracy by speaking like a caveman.