ziniao-access-policy manages erp web access rules, url configurations, url groups, and account-to-policy bindings via direct api calls. no shortcut commands provided, all operations route through ziniao-cli api.

arboreto infers gene regulatory networks from transcriptomics data using grnboost2 and genie3 algorithms, scaling from single machines to clusters. incomplete documentation limits practical deployment.

apify-influencer-discovery provides a multi-platform actor selection framework for finding influencers on instagram, facebook, youtube, and tiktok. execution relies on external apify actors and mcpc cli tooling, with output formatting options but limited decision logic or authenticity verification.

tavily wraps the tavily search api for web research, content extraction, and site crawling. returns structured results with summaries and citations for llm-driven workflows.

uptime-monitoring provides foundational patterns for health checks, status pages, and incident tracking. covers nodejs and kubernetes implementations with best practices guidance but lacks concrete decision trees and failure recovery steps.

stride-analysis-patterns provides categorical threat taxonomy and a decision matrix for systematic security review, but lacks step-by-step execution guidance and defers concrete procedures to external references.

defuddle extracts clean markdown from web pages via cli, stripping navigation and ads to reduce token overhead. prefer over webfetch for standard content extraction.

llm-monitoring-dashboard provides a tokuin-based framework for tracking api costs, tokens, and latency across users with automated dashboard generation. instruction density is high but procedure execution clarity suffers from incomplete examples and missing output validation checkpoints.

presentational-container-pattern teaches separation of view and business logic in react by splitting components into data-fetching containers and pure presentational renderers, with modern hooks as an alternative approach.

sf-deploy covers salesforce deployment orchestration including dry-run validation, manifest-based deploys, and ci/cd workflow guidance. lacks concrete procedural steps and explicit failure handling.

env-and-assets-bootstrap prepares conda environments and asset paths for deep learning repo reproduction, translating readme setup steps into conservative environment assumptions before execution.

adaptive-communication detects conversational patterns (hedging, open framing, personal context) to route between exploratory and transactional responses. clarifies intent when ambiguous before proceeding.

api-documenter covers openapi authoring, developer portal design, and sdk generation. skill is broad but lacks concrete procedural steps, treating most tasks as capability lists rather than actionable workflows.

prompt-caching covers three LLM caching strategies (native prompt caching, response caching, CAG) to reduce token costs and latency, with guidance on invalidation patterns and anti-patterns.

plan decomposes a BDD intent issue into trackable vertical slices grouped into waves, validating write scope isolation and integration order before producing markdown plans and bd issues.

team-rituals guides design of repeatable organizational ceremonies using a named-template-adoption framework. provides cultural principles but lacks concrete procedural depth for implementation.

agile-workflow orchestrates git-based development sequences with checkpoint control, pausing at decision points. covers standup, task progression, sprint ceremonies, and work resumption across platform-agnostic repositories.

xiaohongshu provides a routing layer for automating xiaohongshu operations via mcp. it documents intent classification and global constraints but lacks detailed step-by-step procedures and concrete error handling.

vercel-functions covers serverless and edge compute on vercel with runtime details, cold start timings, and payload limits. lacks deployment workflow and error handling patterns.

visa-doc-translate automates ocr and translation of visa application documents into bilingual pdfs. it handles image rotation via exif, tries multiple ocr backends, and formats output suitable for official submissions.

content-creator provides platform-specific templates and engagement formulas for blog posts, social media, and newsletters. lacks decision logic for audience segmentation and doesn't address failure modes like low engagement or platform algorithm changes.

claude-code-guide documents configuration patterns and debugging tactics for using claude code within projects, centering on claude.md setup, thinking keywords, and context management practices.

release-skills automates multi-language release workflows with semantic versioning and conventional commit parsing. supports node.js, python, rust, claude plugins, and generic projects. includes dry-run mode and optional configuration via .releaserc.yml.

cloudflare-kv exposes edge key-value storage with sub-5ms latency, bulk operations, and ttl support across 1000+ namespaces. eventual consistency within ~60 seconds and 1 write/sec rate limits require retry logic.

youtube-downloader wraps yt-dlp to fetch videos in configurable quality and format. supports quality tiers from 360p to 1080p, multiple output formats (mp4, webm, mkv), audio-only extraction, and custom output paths. defaults to best available quality as mp4.

startup-pivoting applies the four ps framework to help founders diagnose when and how to pivot, emphasizing rational distance from current ideas and appropriate magnitude of change, guided by product leader insights.

agent-orchestrator-task coordinates complex objective decomposition, execution planning, and result synthesis across multiple agents using parallel and sequential strategies with memory-backed progress tracking.

openclaw-control-center provides a unified dashboard for monitoring openclaw agent activity, token consumption, task execution, and cross-session collaboration with security-first permission defaults.

render-deploy handles git-backed service deployments to render via blueprint generation or direct mcp tool creation, with fallback guidance for docker image workflows requiring dashboard access.

huashu-design recommends design philosophy directions from five schools, generates visual demos via agent teams, and delivers critique reports. covers pre-design direction and post-design review workflows with structured 20-style taxonomy.

docx handles creation, reading, and editing of word documents with xml-level control and tracked changes support, but procedures remain incomplete and edge cases largely unaddressed.

distributed-tracing covers implementing jaeger and tempo for request visibility across microservices. includes best practices for sampling, context propagation, and overhead monitoring with basic troubleshooting patterns.

flow-nexus-neural covers neural network training and deployment across single nodes and distributed E2B sandbox clusters, with support for multiple architectures, template marketplace integration, and federated learning. actual capability depth and production readiness remain unclear.

okx-task-watch provides long-poll based task inbox monitoring with multilingual trigger phrases, but lacks step-by-step procedures, error handling, and clear output contracts.

renaissance-architecture provides a philosophy for first-principles design choices across architecture, infrastructure, and ui/ux. it emphasizes earned complexity, local-first defaults, and human-legible systems over cargo-cult adoption of patterns. primarily a reference framework, not a procedural skill.

asc-shots-pipeline automates ios screenshot capture via xcodebuild and simctl, applies ui automation with axe, composes frames using koubou, and uploads to app store connect. experimental locally.

typescript-e2e-testing provides workflow references for jest and docker-based integration testing, but lacks actionable procedures, specific trigger phrases, and failure mode guidance.

project-workflow-analysis-blueprint-generator produces architecture-aware documentation by detecting tech stacks and workflow patterns, then outputting end-to-end implementation blueprints with code examples and extension guidance.

hybrid-search-implementation covers combining vector and keyword search via fusion methods like RRF and cross-encoders, with best practices for tuning and reranking. concrete templates live elsewhere.

gpt-image wraps openai's gpt-image-2 model for text-to-image, editing, and inpainting via belt cli. covers multiple formats, batch generation, and custom resolutions but lacks error handling guidance and intent clarity.

clerk routes authentication tasks to specialized skills based on framework and sdk version, with version detection from package.json and inline callouts for core 2 vs current differences.

business-logic-vulnerabilities outlines a playbook for identifying scanner-invisible flaws in workflows, race conditions, and state machines. relies heavily on external companion files for actionable procedures and specific attack techniques.

opencli-explorer covers api discovery, authentication strategy selection, and adapter authoring via browser devtools. procedurally detailed but primarily assumes bilibili-like platforms and lacks clear trigger phrases for varied use cases.

popup-cro provides strategic frameworks for popup design, trigger selection, and copy patterns. emphasis on timing, value clarity, and respect. lacks procedural depth and concrete implementation guidance.

ads-create generates multi-platform campaign concepts and copy from validated brand inputs, enforcing policy compliance and separating factual claims from creative hypotheses. requires human review before launch.

captions-overlay establishes a doctrine for layering captions as composited overlays atop video content without reserving screen real estate. it defines three caption modes (drop, rail, embed) with strict spacing rules and forbids layout shifts to accommodate text.

diagram-creator generates text-based diagrams in mermaid, plantuml, and ascii. covers flowcharts, sequences, architecture, erd, class diagrams, and timelines. output is code only, not rendered images.