identity-aiops covers keycloak and authentik operations: 27 mcp tools for user/client/event queries, four flagship rca analyses, and governed writes (disable/enable/rotate) with local audit, policy enforcement, undo recording, and encrypted secrets.

showmeai provides agents deterministic image, video, 3d, and audio generation through a python runtime with mandatory readiness checks, state recovery, and credential isolation.

buildin-skills covers the buildin cli for v2 api work including pages, blocks, databases, search, markdown content, and file uploads, with mandatory authentication preflight and explicit write-safety confirmation gates.

alibabacloud-cleversee-search wraps alibaba cloud's web-search cli for ai-driven internet retrieval with optional date-range and domain filtering. requires pre-installed cli binary and valid aliyun credentials via profile system.

aiq-research routes deep research and chat queries to a local or self-hosted nvidia ai-q blueprint backend via python helper script, handling async job polling, report retrieval, and follow-up refinement without managing credentials.

llm-agent builds conversational or tool-calling agents on runware-hosted models via openai-compatible endpoints. the model reasons about user functions, decides when to invoke them, and answers using returned data. covers the standard chat-completions loop with explicit iteration bounds and tool-call pairing.

restore-and-upscale handles deblur, denoise, dehaze, and super-resolution (2x-4x) on degraded images by routing damage type to the right model (real-esrgan, swinir, clarity, ccsr, or bria fibo edit). no dedicated face restoration exists; general upscalers improve faces as a side effect only.

pixel-asset-compiler transforms ai-generated sprite sheets into deterministic game assets through structured validation, compilation, and engine-specific export. handles frame consistency audits, background removal, and godot scene generation with explicit failure recovery.

trade-show-fit-score wraps the lensmor api to generate a quantified exhibit recommendation for a single named event, with explicit score bands, breakdown interpretation rules, and routing to complementary skills for discovery and budget planning.

logos-and-vectors generates or converts raster images to genuine scalable svg with editable paths. covers two workflows (generation via prompt, conversion from png/jpg) across three models, with specific guidance on palette locking, source preparation, and quality verification.

flowus-skills provides structured cli-first workflows for flowus v2 api tasks including pages, blocks, databases, and file operations, with explicit preflight authentication checks and write-safety confirmation gates.

product-ad-video converts a product still into a polished short ad or promo clip via image-to-video generation, with optional product-swap capability for existing footage. targets flagship platforms (social, landing pages, paid launch) and includes draft-to-finish workflow, cinematic prompt guidance, and brand-safety guardrails.

api-and-interface-design covers REST and GraphQL endpoint design, type contracts, and module boundaries. emphasizes hyrum's law, contract-first thinking, consistent error semantics, and backward compatibility patterns.

agent-browser-tekken provides comprehensive browser automation via command-line interface, covering navigation, element interaction via refs, screenshots, network mocking, and parallel sessions. covers form filling, data extraction, web testing, and authentication workflows.

web-accessibility covers wcag 2.1 implementation across semantic html, keyboard navigation, aria attributes, color contrast, and automated testing. provides step-by-step procedures with react code examples for forms, modals, dropdowns, and tabs.

netmiko-ssh-automation covers safe python patterns for read-only network device collection via ssh, bounded batch operations, textfsm parsing, and guarded configuration changes with explicit timeout and error handling.

receiving-code-review covers technical evaluation of feedback before implementation, emphasizing verification and reasoned pushback over performative agreement. applies context-aware rigor to distinguish between trusted and external sources.

ai-assisted-performance-review separates human signal from tooling signal in reviews where ai output dominates, providing criteria rewrites, calibration rules for mixed adoption, and conversation scripts for three hard cases.

skill-haccp-plan-generator automates 11-column haccp plan creation for food safety workflows. it parses production processes, identifies hazards and critical control points, sets measurable limits, and exports compliance tables with explicit marking of default vs user-confirmed values.

account-plan generates structured customer expansion documents with relationship mapping, opportunity quantification, and 90-day execution plans. targets key account strategists and account managers building retention and growth cases.

delay-claim-letter drafts construction delay notices with contract clause citation, cause classification on the excusable-compensable matrix, critical-path impact narrative, and records-preservation guidance. output is notice-ready and counsel-routable.

ai-eval-plan guides design of repeatable evaluation frameworks for llm features, moving from fuzzy quality goals to gated, measurable tests with explicit rubrics, baselines, and regression gates.

decision-memo structures leadership asks around a front-loaded recommendation, decision framing, options table, assumption surface, and explicit deadline. covers reversibility and anti-patterns to avoid.

devils-twin generates a full-strength opposing argument memo before a document ships, showing which claims hold, dent, or collapse under real opposition premises. outputs the counter-memo, claim-by-claim battle map, and a pre-emption paragraph to strengthen the original.

skill-complaint-quality-analysis performs standardized complaint data reviews with multi-sheet parsing, field mapping, statistical analysis, and dual-format report generation. covers complaint ledger analysis, quality problem retrospectives, and anomaly list reviews with explicit workflow separation between agent interaction and script execution.

claims-triage structures first-pass insurance claim assessment across coverage, severity banding, fraud indicators, reserve ranges, and routing recommendations. outputs a triage note with sla and concrete next actions, built for fnol and fast-path decisions.

async-decision-memo runs a bounded, role-based process for remote decision-making: memo structure optimized for silent reading, named decider plus consulted/informed roles, enforced comment protocol, and a closing note that records dissent and the reopen condition. replaces drawn-out comment threads and stalled decision meetings.

ai-product-canvas structures decisions for ai-powered products using a seven-section canvas covering problem definition, model selection, data audit, evaluation metrics, ux design, responsible ai checklist, and launch monitoring. enforces rigor through anti-pattern checks and quality gates.

browser-testing-with-devtools bridges static analysis and live browser execution by giving agents devtools access to inspect dom, read console logs, analyze network requests, and capture performance data. covers ui debugging, network diagnosis, and performance profiling with security boundaries enforced.

candidate-scorecard converts unstructured interview notes into evidence-based competency ratings and hire recommendations, with explicit confidence levels and open questions for calibrated hiring decisions.

database-migration-plan produces a structured four-phase expand/contract migration guide with explicit SQL, backward compatibility analysis, batch backfill procedures, validation queries, and a deployment checklist for zero-downtime schema changes.

remotion-video-pipeline orchestrates audio-driven short-form video production from locked script through CapCut handoff, with two mandatory human gates and delegated Remotion composition work to preserve determinism and prevent rework cascades.

letagentpay enforces deterministic spending policy controls for ai agents, routing all purchase requests through approval workflows, budget enforcement, and crypto-micropayment (x402) authorization with full audit trails.

drizzle covers LobeHub's ORM patterns for pgTable schemas, foreign keys, indexes, type inference, and query builders. emphasizes strict naming conventions, rejects relational APIs in favor of explicit select joins, and documents common pitfalls around enums, jsonb contracts, and id generation.

dropship-evaluator provides a numbered workflow for screening product ideas, evaluating suppliers, modeling margins, and stress-testing communication before committing ad spend. includes decision thresholds, worked rejections, and common failure modes.

refinement-loop covers iterative generate-critique-revise patterns for opus 4.8, with thinking-as-critic modes, model routing cost controls, and concrete convergence detection strategies.

scholar-sidekick-mcp resolves scholarly identifiers (DOI, PMID, arXiv, ISBN, etc.) to structured metadata, formatted citations, and export files. verifies retraction status, open-access availability, and citation authenticity via MCP server integration with RAPIDAPI key.

ln-783-container-launcher builds docker images, orchestrates container startup with dependency ordering, and validates service health via native checks or http probes with exponential backoff retry logic.

find-seo-agency queries the servicegraph api to discover, filter, and enrich us seo agencies by location, sub-specialty (technical, local, ecommerce), quality signals, and third-party credibility. covers direct requests and indirect intent patterns.

pullboard keeps complex multi-agent builds on a shared board instead of agent memory, with work ordered by priority and dependencies, owned by single agents, and verified by different principals to ensure done is earned not asserted.

keystone provides structured api access to document ingestion and hybrid search across knowledge bases, with explicit credential handling, multipart upload support, and polling patterns for async parse operations.

dbs-script-flow diagnoses short-form video scripts for audience drop-off risk by scanning paragraph transitions, information density, and delivery flow. identifies where viewers will swipe away and offers marked-up revisions.

claw-alarm manages alarms on ios via the clawalarm cloud-sync api, supporting create, read, update, delete, and snooze operations. authentication uses a one-time pairing handshake with per-device tokens. all interaction flows through a curl wrapper that auto-discovers endpoints from the live openapi spec.

healthcare-cdss-patterns provides three core modules for building patient-safety-critical clinical decision support: drug interaction checking with severity classification, dose validation accounting for weight/age/renal function, and NEWS2 clinical scoring. emphasizes pure functions, testability, and blocking unsafe prescriptions.

remix-v2-routing covers flat-routes v2 file naming conventions, nested layouts, dynamic segments, resource routes, and decision trees for choosing between layout patterns. concrete examples anchor each pattern.

evaluate-agent-native applies five hard criteria and a url onboarding bonus signal to rigorously classify services as agent-native, agent-adapted, or agent-builder. includes decision tree, evaluation template, and borderline case guidance.

test-driven-development enforces writing failing tests before implementation code, with explicit red-green-refactor cycles. covers the iron law of tdd, common rationalizations, and verification checklists for features and bug fixes.