The library

helpers · When implementing rules—avoids code duplication and ensures AST compatibility.

self-audit · When you need deterministic validation of references without manual verification.

liam · When you need architecture planning or maintenance diagnostics with human-readable output.

synthesize-learnings · When you want to close the loop from analysis findings into meta-skill improvements.

update-ecosystem-index · When you need systematic analysis and actionable recommendations.

sheets-cli · When you need programmatic control of Google Sheets without leaving the terminal.

codegraphcontext · When you need precise AST context for code generation or analysis tasks.

Chart Visualization · When you need to communicate data insights through visual formats.

diatagma · When you need to model complex schemas and generate diagrams systematically.

agt-security · When you need compliance validation against predefined security standards.

agy-cli-agent · When you need agentic task execution with deterministic output contracts.

combine-code-review · Reviewing Combine publishers/subjects when memory leaks or operator chains are suspect

antigravity-project-setup · New projects needing modular agent configuration without monolithic context

claude-cli-agent · Routing specialized analysis tasks without loading full primary agent

healthkit-code-review · Reviewing HealthKit authorization flows and query correctness

claude-project-setup · Optimizing Claude Code setup for team repos with scoped rules

codex-cli-agent · Delegating code analysis to GPT-5 Codex without primary agent overhead

copilot-cli-agent · Isolated analysis tasks via Copilot CLI with controlled model routing

generating-smart-commits · Automating conventional commit messages from staged changes

gemini-cli-agent · Fast, cost-efficient analysis via gemini-3-flash-preview

local-llm-bridge · Sub-second bounded tasks using local Gemma without cloud latency

x402-payment · Requiring payment for API access without traditional billing

local-llm-setup · Day-1 bootstrap or reconfiguration of local LLM inference

clean-architecture · Enforcing strict layering and dependency rules in domain-driven designs

maf-adapter · Testing MAF skills before deployment to production runtime

700-clear-framework_08305a60 · Rapid iterative improvement of prompts and PRD clarity

project-setup · Projects needing both Claude Code and Google ADK configuration simultaneously

dependency-management · Responding to Dependabot security alerts with locked-file workflow

adr-management · Maintaining institutional memory of significant architectural choices

app-router · Building multi-segment routes with dynamic parameters and layouts

coding-conventions-agent · Enforcing dual-layer documentation and naming consistency across languages

auth-patterns · Authentication-first Next.js apps requiring OAuth or custom credential handling.

context-bundler · Sharing large codebases with external agents while respecting context limits.

route-handlers · Building REST APIs in Next.js with streaming and proper error handling.

using-gh-cli · Accessing private repos without WebFetch, leveraging automatic user authentication.

convert-mermaid · Rasterizing diagrams to images when live rendering is unavailable.

server-actions · Mutating data from client-side forms with server validation and cache revalidation.

hf-download · Fetching HuggingFace models with automatic exponential backoff on rate limits.

server-components · Reducing client JavaScript while enabling direct database access via RSC.

2700-scan-secrets_2084d9d6 · Detecting hardcoded API keys before committing to version control.

hf-init · One-time setup of HuggingFace datasets with standard ADR 081 structure.

refs · Standardizing skill authorship with validated frontmatter and boilerplate.

hf-upload · Persisting agent learnings and semantic caches to HuggingFace with tags.

humanize · Removing LLM fingerprints to sound authentically human across channels.

git · Version control operations without native language support.

link-checker-agent · Validating documentation and content links before deployment.

037-memory_space_edits_729e16d3 · Persisting project context and preferences across Claude Code sessions.

optimize-context · Reducing token usage while preserving essential information in long conversations.

repository-inference · Automatically detecting project type and structure without explicit config.

red-team-bundler · Preparing code for adversarial testing without exposing sensitive data.