Agent Skills by xoai
Showing 54 vetted skills indexed across 1 GitHub repositories.
session-bridge
Restore agent session context from plan file checkboxes.
tool-use
Select and execute external tools across local scripts, MCP proxy, and subagent dispatch layers.
scope-guard
Monitor code changes to enforce implementation scope and detect deviations.
context-loader
Define three content loading layers for AI agent context windows.
verify-completion
Execute test suites and commands to verify code implementations against acceptance criteria.
systematic-debug
Implement a four-phase debugging framework for root cause analysis.
plan
Generate implementation plans with file paths, tests, and dependency ordering.
specify
Convert elicitation output or PRDs into formal specifications defining what to build and why.
quick-elicit
Guides a three-round questioning process to elicit feature specifications with codebase-scan context.
codebase-scan
Scan codebase areas to identify patterns, dependencies, and architecture.
deep-elicit
Conduct Socratic conversations to elicit product requirements and produce a brief.
implement
Implement a planned task by writing tests, minimal code, and refactoring.
tdd
Enforce a RED-GREEN-REFACTOR cycle for production code implementation.
quality-review
Evaluate code quality, security, maintainability, and performance beyond specifications.
spec-review
Compare code implementation against task specification for completeness and precision.
visual-review
Analyze UI screenshots at multiple breakpoints against design specifications and web standards.
sage-navigator
Assess project state and recommend proportional next process steps.
onboard
Automate Sage AI agent framework setup by detecting tech stacks and generating configuration files.
build-loop
Execute approved plans task-by-task with quality gates between steps.
sage-help
Analyze project state to provide next-step guidance in the Sage framework.
opportunity-map
Generate an opportunity map by assessing customer needs against product capabilities.
ux-evaluate
Compare a design system against category benchmarks to produce a structured gap analysis.
user-interview
Design user interview research packages with briefs, screeners, guides, and analysis frameworks.
pack-source-process
Gather and filter sources for AI agent pack building.