By mithro
Blocks SSH hostname hashing (-H flag and HashKnownHosts=yes) — keeps known_hosts readable
npx claudepluginhub mithro/dot-claude --plugin hook-block-ssh-hash-hostnamesPersonal Claude Code plugin marketplace.
claude plugin marketplace add mithro/dot-claude
Then enable plugins per-machine:
claude plugin enable agent-debugger@mithro-personal
claude plugin enable hook-block-tmp-creation@mithro-personal
claude plugin enable my-conventions@mithro-personal
# etc.
Specialized agents for various development tasks:
| Plugin | Description |
|---|---|
agent-accessibility-tester | WCAG 2.1 compliance, semantic HTML, ARIA, keyboard nav |
agent-api-designer | REST API design, DRF, OpenAPI, versioning |
agent-backend-architect | Scalable API design, microservices, distributed systems |
agent-celery-expert | Async task debugging, retry strategies, queue management |
agent-code-reviewer | Code quality, security vulnerabilities, best practices |
agent-data-scientist | SQL optimization, Django ORM, data visualization, Pandas |
agent-debugger | Root cause analysis, systematic debugging, profiling |
agent-deployment-engineer | Production deployment, zero-downtime, WSGI/ASGI |
agent-devops-engineer | CI/CD, Docker, GitHub Actions, infrastructure as code |
agent-django-developer | Django 5.2+, REST APIs, async views, Celery |
agent-documentation-writer | API docs, docstrings, architecture docs |
agent-error-detective | Error patterns, stack traces, Sentry integration |
agent-performance-engineer | Profiling, caching, async performance, optimization |
agent-postgres-pro | PostgreSQL 17, JSONB, full-text search, query tuning |
agent-python-pro | Modern Python 3.11+, type safety, async programming |
agent-security-auditor | OWASP Top 10, dependency scanning, security review |
agent-solution-researcher | Multi-approach evaluation, trade-off analysis |
agent-sre-engineer | Monitoring, observability, incident response, SLOs |
agent-test-specialist | Django/pytest, browser testing, test coverage |
Enforcement hooks for SSH, git, and command safety:
| Plugin | Description |
|---|---|
hook-allow-ssh | Auto-allows SSH/scp/rsync to configured hosts (eliminates permission prompts) |
hook-block-tmp-creation | Blocks file creation in /tmp/ (use project-local tmp/) |
hook-block-stderr-to-null | Blocks 2>/dev/null (keep diagnostic output visible) |
hook-block-ssh-hash-hostnames | Blocks SSH -H flag and HashKnownHosts (keep known_hosts readable) |
hook-block-ssh-unsafe-options | Blocks StrictHostKeyChecking=no and UserKnownHostsFile=/dev/null |
hook-block-inline-scripts | Blocks python -c and heredocs (write to a file first) |
hook-safe-force-push | Blocks bare git push --force (use git safe-force-push <branch>) |
| Plugin | Description |
|---|---|
my-conventions | Personal coding conventions: Python/uv, ISO dates, small commits, Apache 2.0, etc. |
github-repo-setup | GitHub repository configuration skill and tag ruleset script |
dot-claude/
├── .claude-plugin/
│ └── marketplace.json # Marketplace catalog (28 plugins)
├── marketplace/
│ ├── agent-*/ # 19 agent plugins
│ ├── hook-*/ # 7 hook plugins
│ ├── my-conventions/ # Coding conventions (SessionStart hook)
│ └── github-repo-setup/ # GitHub config skill + scripts
├── settings.json # Repo-level permissions
├── CLAUDE.md # Repo description
├── GitHub.md # Reference: GitHub setup commands
└── LICENSE # Apache 2.0
Licensed under the Apache License, Version 2.0. See LICENSE for the full license text.
Tim 'mithro' Ansell
Ultra-compressed communication mode. Cuts ~75% of tokens while keeping full technical accuracy by speaking like a caveman.
Executes bash commands
Hook triggers when Bash tool is used
Modifies files
Hook triggers on file write and edit operations
Intelligent prompt optimization using skill-based architecture. Enriches vague prompts with research-based clarifying questions before Claude Code executes them
Persistent memory system for Claude Code - seamlessly preserve context across sessions