Skip to content

Repository files navigation

AI agents skills set

A set of AI agent skills designed to solve research and development tasks.

Skills Overview

Research

General-purpose research and ideation skills. Each skill is independent and addresses a different research intent — from open-ended exploration to structured ideation and in-depth synthesis.

Skill Description
deeper-research Conduct thorough, multi-source research using Tree of Thoughts methodology. Produces cited, multi-perspective syntheses with confidence ratings and evidentiary hierarchy.
explore Explore → Plan → Implement — a thinking partner for exploring ideas and clarifying requirements, then producing a structured roadmap (What/How/Where/Acceptance criteria) for approval, then implementing it.
scamper Transform existing ideas into improved variants using the SCAMPER method (Substitute, Combine, Adapt, Modify, Put to other use, Eliminate, Reverse).
triz-solver Solve inventive engineering, product, process, or system-design problems using TRIZ methodology and contradiction resolution.
vibe-research Iterative research copilot for exploratory "vibe research" — progressively sharpens broad topics into high-value lines of inquiry.

Engineering Research

Skills that automate the Iterative Engineering Research Methodology — from initial brief to final report. See engineering/README.md for the full workflow.

Skill Description
research-init Bootstrap a new research project: create the directory structure, populate the brief, and register the project in the index.
research-prior-art Search for and catalog existing work — papers, CVEs, tools, talks, blogs, standards — with relevance ratings.
research-hypothesis Create, update, and manage hypothesis cards and the hypothesis graph (Mermaid DAG + catalog).
research-experiment Guide experiment design, execution tracking, and result recording for a research hypothesis.
research-decision Analyze hypothesis results and recommend continue, pivot, kill, or fork decisions.
research-synthesis Execute the synthesis phase: determine report mode, walk the hypothesis graph, and generate the final report.
research-status Generate a comprehensive status overview of a research project.

Agentic

Skills for building and maintaining AI agent systems.

Skill Description
prompts-review Review all prompts in a codebase for optimality, balancing effectiveness and token efficiency. Covers explicit, string-literal, and dynamically constructed prompts.

Security

Skills for security analysis and policy generation.

Skill Description
security-policy-generator Analyze a project repository and generate a comprehensive SECURITY.md with threat model, security architecture, secure coding guidelines, and OWASP Top 10 for Agentic Applications (ASI01–ASI10) coverage.

Writing

Skills for text transformation and post-editing of AI-generated content.

Skill Description
humanize-ru Remove signs of machine generation from Russian-language AI text while preserving its original register and communicative purpose — scientific stays scientific, business stays business.

Development

Skills for writing secure, idiomatic, production-ready Go applications — from specification to correct code.

Skill Description
vibe-brain Analyzes the project codebase, documentation, and memory to identify the single highest-priority task to work on right now.
code-review Comprehensive code review for evaluating code changes — identifies bugs, logic errors, security issues, structural problems, performance concerns, and unintended behavior changes.

Secure Go

Skill Description
secure-go Write secure Go applications through everyday practices — idiomatic patterns, standard library defaults, OWASP Top 10 coverage, and a security-focused linter setup. Covers access control, data protection, input handling, secure design, configuration, dependencies, authentication, error handling, integrity, logging, external objects, and SSRF.

Specification-Driven Development

Skills that implement a Specification-Driven Development workflow — from bootstrapping a spec system to keeping specs aligned with code. See development/sdd/README.md for the full workflow.

Skill Description
vibespec-init Initialize a specification system for a project from scratch by analyzing the codebase and generating the initial set of specs.
vibespec-create Create a new specification document following the project's spec system templates and conventions.
vibespec-consult Proactively consult project specifications before making structural changes to avoid violating documented invariants.
vibespec-update Update existing specification documents after code changes that alter documented behavior, interfaces, or invariants.
vibespec-check Detect discrepancies between project specifications and actual code, interactively resolving each finding with the user.

Idiomatic Go

Skills for writing idiomatic, correct, and performant Go code. See development/idiomatic-go/README.md for how the skills are organized and when to use which.

Skill Description
go-code-organization Well-organized Go code and projects — variable scoping, nested control flow, and project structure.
go-concurrency-foundations Foundational concurrency concepts — concurrency vs parallelism, goroutine scheduling, and channel semantics.
go-concurrency-practice Practical Go concurrency rules — goroutines, channels, mutexes, and synchronization patterns.
go-control-structures Correct usage of range loops, break statements, and defer in loops.
go-data-types Common mistakes with basic data types, slices, and maps.
go-error-management Go error management — panic usage, error wrapping, and checking error types and values.
go-functions-methods Best practices for Go functions and methods — receivers, named results, nil receivers, and defer.
go-optimizations Performance-sensitive Go code — CPU cache, escape analysis, allocation reduction, profiling, and GC tuning.
go-standard-library Common Go standard library mistakes — time.Duration, JSON handling, and HTTP pitfalls.
go-strings Correct and efficient Go string code — iteration, concatenation, and conversions.
go-testing-mistakes Common Go testing mistakes — test categorization, race detection, and test execution.

Quick Start

Installation

These skills can be installed using the Skills CLI from skills.sh:

# Install all skills collection
npx skills add v0lka/skills --all --full-depth
# Install a specific skill
npx skills add v0lka/skills --skill <skill-name> --full-depth

Manual Usage

To use a skill manually, reference the agent skills documentation or load it via your AI agent interface.

License

MIT License — see LICENSE for details.

About

A set of AI agent skills for research and development tasks.

Topics

Resources

Stars

15 stars

Watchers

0 watching

Forks

Releases

Packages

Contributors

Languages