Kiwi AI
Official@kiwi-home
Offers structured codebase analysis, project governance, and multi-specialist coordination protocols for enterprise software development lifecycles.
Agent Skills by Kiwi AI
Showing 24 vetted skills indexed across 2 GitHub repositories.
agent-discovery
Scans .claude/agents/*.md and ranks specialists by domain overlap and trigger context.
quality-gates
Enforce structured quality gates across pipeline stages using workflow.yaml.
context-resolution
Resolve project workflow context and git push targets from repository configuration files.
environment-detection
Detect runtime and tool configuration for Python, Node, Ruby, Rust, or Go codebases.
adversarial-review
Analyze plans across six risk vectors and produce severity-tiered findings.
questionnaire
Collect project setup preferences and map answers into workflow.yaml and quality-gate configuration.
feedback-loop
Automate remediation iterations for blocking review findings with retry limits.
context-file-generation
Generate subdirectory CLAUDE.md files with provenance markers and workflow.yaml reconciliation entries.
brief-writer
Convert one-sentence intents into structured work briefs with YAML frontmatter and acceptance criteria.
plan-writer
Generate structured implementation plans from design session outputs and briefs.
deliverable-review
Evaluate delivered artifacts against the brief with severity tiers and disposition rules.
asset-reconciliation
Detect, classify, and reconcile generated and manual project assets during setup re-runs.
domain-detection
Detect project domain from repository file and directory signals.
stack-detection
Identify languages, frameworks, and domains in codebases via mapping tables.
agent-patterns
Document agent frontmatter metadata spec and discovery patterns for .claude/agents/*.md files.
deliberation-protocol
Coordinate multi-round specialist deliberations with conflict detection and escalation paths.
tdd-patterns
Identify stack-aware TDD patterns and anti-patterns for test design.
systematic-debugging
Guide hypothesis-driven debugging through a five-phase failure resolution process.
agent-team-protocol
Coordinate parallel code execution teams with TDD workflows and Git strategy.
codebase-analysis
Analyze codebases to reveal patterns and conventions for agent and skill design.
issue-writer
Generate requirements-focused issues with acceptance criteria and constraints.
asset-discovery
Map existing skills and agents across project, user, and plugin layers.
issue-workflow
Plan and execute GitHub issues with research, requirements extraction, and build-vs-buy evaluation.
pr-review
Classify PR review findings by severity tiers for merge decisions.
Frequently Asked Questions About Kiwi AI
FAQPage SchemaWhat specific tasks can I perform using Kiwi AI's skill manifest?▼
You can perform automated codebase analysis, generate structured implementation plans from briefs, conduct adversarial reviews across risk vectors, and reconcile project assets. These skills enable systematic debugging, requirement extraction, and the enforcement of quality gates within your repository configuration files.
Which personas benefit most from these development protocols?▼
Software architects, lead engineers, and technical project managers benefit from these protocols. The skills are designed for teams requiring rigorous documentation, standardized issue generation, and structured review processes to maintain consistency across complex, multi-specialist development environments.
What are the prerequisites for implementing these project governance skills?▼
Implementation requires a repository structure that supports .claude/agents/ metadata files and standard YAML configuration files. The system relies on existing repository signals, such as directory structures and configuration files, to detect project domains, runtime environments, and stack-specific TDD patterns.