A collection of language-agnostic coding-agent skills for code quality and architecture. They use the open SKILL.md format and can be installed for Codex, Claude Code, Cursor, GitHub Copilot, OpenCode, and other supported agents.

Architecture review based on semantic design costs rather than language-specific syntax.

/decomplect

• Simplicity - Separate independent concerns and expose hidden context
• FCIS - Separate deterministic policy from effects where it pays off
• Coupling - Align cohesion, boundaries, and dependency direction

View skill source →

Evidence-first tactical review of contracts, responsibilities, failures, and duplication.

/unslopify

• Contract Strength - Express and enforce valid states idiomatically
• Responsibility - Keep independent change vectors separate
• Failure Integrity - Make recovery, degradation, and failure explicit
• DRY - Remove duplicated knowledge without premature abstraction

View skill source →

Install every skill globally for all supported coding agents with the Vercel Labs Skills CLI:

npx skills add shanev/skills --all -g

Or install one skill globally for all supported agents:

npx skills add shanev/skills --skill decomplect --agent '*' -g -y
npx skills add shanev/skills --skill unslopify --agent '*' -g -y

Omit -g to install into the current project instead of your user-level agent directories. The CLI uses symlinks by default so each agent shares one canonical copy of a skill.

Claude Code users can alternatively install the skills as plugins:

/plugin marketplace add shanev/skills

# Install what you need
/plugin install decomplect@shanev-skills
/plugin install unslopify@shanev-skills

Verify installation:

/plugin

Ask your coding agent to apply a skill by name:

Use decomplect to review the architecture of this branch.
Use unslopify on src/payments and focus on failure handling.

Claude Code plugin users can also invoke the slash commands directly:

/decomplect:decomplect    # Architectural analysis
/unslopify:unslopify      # Tactical cleanup

Both skills accept explicit files, directories, diffs, commits, branches, or whole-repository scopes. They work across programming languages by evaluating behavior and contracts relative to the repository's own ecosystem.