Findings & checks · 0 flagged
Every scanned point with the score it earned and what moved between them.
First recorded scan — no prior version to compare against.
The primary manifest — the file an agent reads to learn what this artifact does.
Skills
Small, composable agent skills for real engineering work. Each skill is independent, trigger-activated, and designed to work with Claude Code and any skills.sh-compatible agent.
Why
Monolithic agent frameworks (GSD, BMAD, Spec-Kit) try to own your entire workflow. They work until they don't — then debugging the framework is harder than doing the work yourself.
These skills take the opposite approach:
- Small — each skill does one thing well (diagram, debug, grill, handoff)
- Composable — skills reference each other when useful, never depend on each other
- Adaptable — fork, modify, keep what works, discard what doesn't
- Model-agnostic — plain markdown instructions, no vendor lock-in
Install
npx skills@latest add TheAkshitS/skillsPick the skills and agents you want. To grab one skill non-interactively:
npx skills@latest add TheAkshitS/skills -s c4-views -ySkills
- [c4-views](./skills/c4-views/SKILL.md) — Generate and review C4 model architecture diagrams in Mermaid, Structurizr DSL, or C4-PlantUML. _Trigger: "diagram", "architecture", "C4", "system context", "container diagram", "component diagram"_
- [external-model](./skills/external-model/SKILL.md) — Run a prompt through a different AI model via the opencode, cursor-agent, or kiro-cli command-line tools. _Trigger: "second opinion", "what would GPT-5/Gemini say", "delegate to cursor/opencode/kiro", "run with model X", "ask another model"_
Skill structure
skills/<skill-name>/
├── SKILL.md # Instructions loaded by the agent (required)
├── AGENTS.md # Skill-specific agent hints (optional)
├── README.md # Standalone distribution docs (optional)
├── references/ # Long-form material loaded on demand (optional)
└── scripts/ # Utility scripts (optional)Skills stay flat under skills/ until 3+ skills exist, then organize into buckets: engineering/, productivity/, misc/. See docs/repo-layout.md for details.
Develop locally
Symlink every skill into your local agent (~/.claude/skills):
bash scripts/link-skills.shValidate frontmatter before committing:
bash scripts/validate-skills.shReleases
Automated via release-please on push to main. See docs/local-workflow.md for the file layout, semver bump rules, and bootstrap notes.
License
MIT — see LICENSE.