Turn your AI assistant into a proactive Senior Engineer.

This repository is not just a collection of "docs"—it is a library of Active Cognitive Skills designed to give your AI agent (Cursor, Windsurf, Claude Code, Gemini) the ability to reason, plan, and execute complex engineering tasks autonomously.

When you install these skills, you aren't just giving your AI context; you are giving it a Role.

Compatible with vercel-labs/skills. Works with 54+ agents including Claude Code, Cursor, Cline, Codex, and more.

# Install all skills into your project
npx skills add xblaster/instill-skills

# Install globally (available across all projects)
npx skills add xblaster/instill-skills --global

# Install specific skills only
npx skills add xblaster/instill-skills --skill git-master --skill dependency-sentinel

# Target a specific agent
npx skills add xblaster/instill-skills --agent claude-code

# List available skills without installing
npx skills add xblaster/instill-skills --list

# Add this repo as a remote source
instill sources add https://github.com/xblaster/instill-skills

# Install the skills you need
instill install dependency-sentinel git-master

• Dependency Sentinel (dependency-sentinel)
  • The Guardian. Safely audits and remediates vulnerabilities. It doesn't just "report" issues—it uses a state machine to verify that updates don't break your build before committing them.
• Security Audit (security-audit-tool)
  • The White Hat. Performs comprehensive security scans using tools like Semgrep or Snyk, creating prioritized remediation plans and ensuring your codebase is fortress-ready.

• Debug Stacktrace Resolver (debug-stacktrace-resolver)
  • The Fixer. Analyze, research, and resolve complex software errors and stack traces. It performs environment discovery, structured web research, and proposes deterministic fixes after verifying context.

• Architect Vision Audit (architect-vision-audit)
  • The CTO. Ensures your code aligns with the project's long-term vision. It detects "drift" between your documentation and your actual implementation, preventing scope creep and architectural decay.
• Tech Debt Loan Officer (tech-debt-loan-officer)
  • The Auditor. Treats code complexity as financial liability. Before you merge that "quick fix," this skill calculates the interest rate on your technical debt and advises on whether you can afford the loan.

• Git Master (git-master)
  • The Release Manager. Orchestrates the perfect commit. Handling everything from conventional commit messages to resolving complex merge conflicts and crafting descriptive Pull Requests.
• Generate Dockerfile (generate-dockerfile)
  • The Ops Expert. Instantly detects your tech stack (Node, Python, Go, Rust) and generates a production-hardened, multi-stage Dockerfile faster than you can say "containerize."

• Auto-Doc Manager (auto-doc-manager)
  • The Librarian. Banish "usage rot" forever. This skill automatically synchronizes your code changes with your documentation, ensuring your READMEs, JSDoc, and OpenAPI specs are always up to date.

• Improbable Recruiter (improbable-recruiter)
  • The Talent Scout. Identifies the "Black Swan" hire. It analyzes your codebase's "Technical DNA" to generate a job description for an atypical profile who would bring disruptive innovation.

• OSS Community Catalyst (oss-community-catalyst)
  • The Evangelist. Audit your repo for "social proof" and generate high-engagement technical content. It identifies friction in your onboarding flow and creates a launch plan to amplify your project's reach.

Each skill in this repository follows a structured, state-machine approach to ensure reliability and safety:

1. Input Handling & Resolution: Context discovery and proactive ambiguity resolution (asking questions instead of hallucinating).
2. Pre-flight & Discovery: Silent verification of the local environment, dependencies, and tool availability.
3. Execution: Running deterministic commands (preferring --json output) to gather ground-truth data.
4. Reporting & UX: Structured Markdown output with clear summaries and actionable next steps.
5. Guardrails: Explicit safety constraints to protect your codebase, secrets, and data integrity.

skills/
├── architect-vision-audit/
│   └── SKILL.md        # Architectural alignment & CTO-level oversight
├── auto-doc-manager/
│   └── SKILL.md        # Documentation synchronization
├── debug-stacktrace-resolver/
│   └── SKILL.md        # Error analysis & resolution
├── dependency-sentinel/
│   └── SKILL.md        # Vulnerability auditing & remediation
├── generate-dockerfile/
│   └── SKILL.md        # Tech stack detection & Dockerization
├── git-master/
│   └── SKILL.md        # Advanced Git workflows & PR management
├── improbable-recruiter/
│   └── SKILL.md        # Strategic hiring & cognitive diversity analysis
├── oss-community-catalyst/
│   └── SKILL.md        # Community engagement & OSS growth
├── security-audit/
│   └── SKILL.md        # Comprehensive security scanning
└── tech-debt-loan-officer/
    └── SKILL.md        # Technical debt & complexity analysis

Each SKILL.md uses YAML frontmatter (name, description) for discovery by npx skills and compatible agents.

Got an idea for a new "Cognitive Module"?

1. Create a directory under skills/ (e.g., skills/my-skill/).
2. Add a SKILL.md file with the required frontmatter:

   ---
   name: my-skill
   description: What this skill does and when to use it
   ---

3. Define the input, steps (the "thought process"), and output. Use our Template Skill as a starting point.
4. Open a PR!

Part of the Instill Ecosystem - Empowering Agentic Workflows.