111 battle-tested AI skills for the complete development lifecycle. From brainstorming to bug bounty to robot deployment, in one place.
This is a curated integration of skills from five sources:
- obra/superpowers — 15 core process skills. The methodology framework: debugging, testing, planning, code review.
- Claude-BugHunter — 52 security skills. Built from real engagements with 681 disclosed vulnerability reports.
- robotics-superpowers — 33 robotics and language skills. ROS, sensor fusion, localization, plus Python, C++, Rust, JavaScript, and more.
- stop-slop — 1 writing skill. Remove AI-generated patterns from prose. Active voice, no filler, no adverbs.
- agentic_installers_tools — 9 skills. Navya autonomous vehicle ecosystem (C++ build, Git flow, RTMaps), skill discovery, and developer tooling.
Total: 111 skills.
Clone it:
git clone git@github.com:RedEarth-Robotics/unified-superpowers.git
cd unified-superpowersOr use the installer:
./scripts/install-universal.shThe installer auto-deploys skills to 7 AI coding platforms: .claude, .opencode, .devin, .codeium, .codex, .copilot, .cursor. Skip with --no-platforms.
Or use the new interactive installers:
# Unified token installer (DCP, skills framework, MCP servers, tools)
bash scripts/unified-token-installer.sh
# Script manager (browse and run all scripts)
bash scripts/scripts-manager.shVerify:
find skills -name "SKILL.md" | wc -l
# Should output: 111| Domain | Count | Highlights |
|---|---|---|
| Security | 52 | 28 hunt modules, bug bounty, pentesting, OSINT |
| Core Process | 17 | Debugging, TDD, planning, code review, parallel agents, framework orchestration, skill discovery |
| Robotics | 18 | ROS, GPS/INS fusion, SLAM, embedded systems |
| Languages | 17 | Python, C++, Rust, JavaScript, MATLAB, DevOps, Linux, likec4, Slidev |
| Navya | 6 | C++ build (Doyle), Git flow, RTMaps, configuration (Grimoire), drive composition (Spellcaster) |
| Writing | 1 | Remove AI slop from prose. Active voice, no filler |
See SKILL_CATALOG.md for the full list.
Skills auto-trigger based on what you ask. No manual loading, no configuration files.
Security example:
"Audit this auth system for vulnerabilities"
→ Activates security-pentester + hunt-auth-bypass
Robotics example:
"My underwater drone loses position when GPS is weak"
→ Activates gps-ins-localization-expert + fusion-filter-robotics-expert
Development example:
"This Python function is slow"
→ Activates python-expert + systematic-debugging
Writing example:
"This README sounds like AI wrote it"
→ Activates stop-slop
Navya example:
"How do I build this C++ project with Doyle?"
→ Activates doyle + navya
Security
bug-bounty— Complete workflow: recon → hunt → reportsecurity-pentester— General pentesting with payloads and bypass tableshunt-xss— Cross-site scripting hunting methodologyoffensive-osint— Open-source intelligence gathering
Robotics
ros-robotics-expert— ROS development, debugging, integrationgps-ins-localization-expert— GPS/INS sensor fusion for localizationrobotics-data-analyzer— Sensor data analysis and diagnostics
Development
systematic-debugging— Structured bug investigation methodologybrainstorming— Design and specification creationpython-expert— Debugging, optimization, best practicescpp-expert— Review, safety, MISRA compliancefind-skills— Discover and install skills from the open ecosystem
Navya
doyle— C++ workspace and build manager (Conan/CMake)navya-git-flow— Git workflow for C++ projects and Chaudron integrationspellcaster— Drive composition and deployment CLI
All scripts support -d (custom directory), -y (automated), --skip-deps, --no-platforms, -v (verbose).
# Universal (cross-platform)
./scripts/install-universal.sh -d ~/skills --yesInteractive installers with menus, dry-run support, and component selection:
# Unified token installer — DCP, skills framework, MCP servers, tools
bash scripts/unified-token-installer.sh # Full auto mode
bash scripts/unified-token-installer.sh --interactive # Menu-driven selection
bash scripts/unified-token-installer.sh --minimal # Core stack only
bash scripts/unified-token-installer.sh --dry-run # Preview without installing
# Script manager — browse and run all project scripts
bash scripts/scripts-manager.sh # Interactive menu
bash scripts/scripts-manager.sh --dry-run # Preview mode# Backup/restore AI configurations
bash scripts/backup-restore-ai-configs.sh save
bash scripts/backup-restore-ai-configs.sh restore --from=~/backups/latest
# Maintain and update stack components
bash scripts/maintain-stack.sh --check # Check for updates
bash scripts/maintain-stack.sh --update # Interactive update
# Run benchmark suite
bash scripts/run-benchmark.shFor full details, see INSTALLATION.md and scripts/README.md.
- SKILL_CATALOG.md — All 111 skills, organized by domain
- docs/ENGAGEMENTS.md — Security skill provenance: real engagement records
- INSTALLATION.md — Platform-specific setup and troubleshooting
- scripts/README.md — Scripts and installer documentation
This is a curated integration. Update from upstreams manually:
cd /path/to/superpowers && git pull # core process skills
cd /path/to/Claude-BugHunter && git pull # security skills
cd /path/to/robotics-superpowers && git pull # robotics & language skillsWe check for upstream updates monthly.
Contribute to the original repositories, not here:
- obra/superpowers — Core process skills
- Claude-BugHunter — Security skills
- robotics-superpowers — Robotics & language skills
- stop-slop — Writing skills
- agentic_installers_tools — Navya ecosystem and developer tooling
This repo is the integration layer. It preserves original skill structure, maintains all original licenses, and handles cross-platform deployment.
See individual skill directories. Each skill retains its original source repository's license.
Built by RedEarth-Robotics by integrating skills from obra, elementalsouls, and RedEarth-Robotics.