architecture-decision-records
A skill for capturing key architectural decisions and their tradeoffs in ADRs.
chaos-readiness
A skill for validating failure modes through safe fault injection.
database-migrations
A skill for managing database schema changes using Entity Framework Code First migrations. Ensures production databases always reflect the migration history with no manual schema changes.
default-web-stack
A skill for providing the default web stack for projects. This dictates which technologies to leverage across projects, unless specified otherwise by the user. This includes backend, frontend, database, security, and so on. When no defaults are provided in this skill and a decision is made, the architect should update this skill accordingly.
documentation
A skill for generating and maintaining project documentation, including design documents, user guides, and API documentation. This skill ensures that all documentation is clear, comprehensive, and up-to-date to facilitate effective communication and knowledge sharing among team members and stakeholders and based on the repository structure.
linting
A skill for enforcing code quality standards through linting across multiple platforms and languages. This skill provides platform-agnostic guidance for configuring, running, and maintaining linters with zero-warning tolerance for .NET, JavaScript/TypeScript, Python, and other platforms.
memory
A skill for managing agent memory and knowledge persistence. Agents should use memory tools by default for storing and recalling information. If memory tools are unavailable, fall back to docs/memories.json with strict JSON validity requirements.
performance-budgeting
A skill for defining and enforcing performance budgets to prevent regressions.
privacy-review
A skill for reviewing data handling, retention, and access to reduce privacy risk.
regression-testing
A skill for providing agent guidance on performing manual regression testing using browser automation tools. The agent must test all pages at 4 breakpoints, all interactive states, and monitor console for errors. Use interactive testing tools only - NOT npm/npx commands or spec files for manual regression.
release-readiness
A skill for ensuring safe and deliberate releases with clear go or no-go checks.
repository-structure
A skill for understanding and adhering to the repository structure. This includes knowledge of where different types of files are located, how the codebase is organized, and best practices for maintaining a clean and efficient repository structure. The agent should be able to navigate the repository, understand the purpose of different directories and files, and provide suggestions for improving the structure if necessary.
self-improvement
A skill for continuous improvement by capturing friction, proposing changes, and verifying impact using Gaia improvement tools.
skill-maturity
A skill for assessing skill coverage, identifying gaps, and prioritizing new skills to add.
spec-consistency
A skill for keeping code and specs aligned in a spec-driven workflow.
start-projects
A skill for understanding how to run projects both locally and in production. This includes knowledge of the necessary tools, dependencies, and configurations required to set up and run the project successfully. The agent should be able to provide clear instructions for running the project in different environments, troubleshoot common issues that may arise during setup, and ensure that the project is running smoothly.
system-evolution
A skill for assessing change impact across components before implementation.
test-strategy
A skill for tailoring test depth to risk and defining coverage targets.
threat-modeling
A skill for identifying security risks using lightweight, repeatable prompts.
unit-testing
A skill for writing comprehensive unit tests across multiple platforms and frameworks. This skill provides platform-agnostic guidance for achieving high test coverage, writing effective test cases, and following testing best practices for .NET (xUnit), JavaScript/TypeScript (Vitest, Jest), Python (pytest), and other platforms.
web-research
Two-tier web research using web fetching as primary for known URLs and headless browser automation as fallback for searches. Requires 3+ sources, prioritizes official docs, includes versions/dates.