startup routes founder decisions across product, growth, hiring, and fundraising by stage and pmf signals. it orchestrates specialized agents, applies stage-appropriate priorities, and surfaces measurable decision gates to avoid common failure modes.

bangkok covers visitor itineraries, neighborhood selection, cost breakdowns, visa strategy (including the 2024 DTV), transport, and nomad infrastructure. routes by user role and timeline rather than averaging answers.

cfo provides financial guidance across forecasting, runway, fundraising, and operations. covers stage-specific advice, escalation rules for human decisions, and explicit traps. modular loading via sub-documents (planning.md, cash.md, fundraising.md, operations.md).

cmo covers marketing strategy, positioning, demand generation, channel selection, and budget allocation across company stages. routes to sub-skills for brand, demand, content, and operations work. emphasizes pipeline math and positioning before channel tests.

test-patterns covers unit, integration, and e2e testing across node.js, python, go, rust, and bash. includes mocking, coverage measurement, fixture setup, and tdd workflow guidance with concrete examples.

codifica provides a git-backed protocol for multi-agent task coordination using plain text state files, execution notes, and atomic claims to maintain persistent context across handoffs.

accountsos provides transaction tracking, vat management, and deadline alerts for uk micro-businesses via api-first accounting infrastructure. handles categorization, invoicing, and director's loan account monitoring for limited companies and sole traders.

idea-generation-and-ideation operationalizes ideation via the spark framework (stock inputs, produce volume, angle to brand, rank and bank, kick-test). it positions proprietary audience signals (comments, dms, faqs) as the discovery engine and ai as the expansion lever, rejecting consensus-model sameness through weekly review cadence and scored idea banks.

ib-suite is a read-only interactive brokers diagnostic toolkit that ingests account data, positions, trade history, and dividends into a local data lake, then produces graded portfolio reports. never places orders.

bana-realestate wraps a chinese property data api covering cities, communities, secondhand homes, rentals and new developments. executes queries via python client, handles credential management in-band, enforces cost disclosure before paid calls, and distinguishes free vs paid methods clearly.

content-audit reads 90 days of native analytics and account output to triage what's working and what isn't, surfacing signal over vanity metrics and feeding a keep/kill/refresh roadmap to downstream skills like content-recycling and profile-optimization.

flow-voice clones voices from 3-30 second audio samples and generates speech locally via luxTTS. runs on cpu and apple silicon, requires no api key, outputs 48kHz wav or video-embedded audio.

web2labs-studio automates video editing for creators, transforming raw recordings into jump-cut videos, vertical shorts, captions, and thumbnails. supports batch workflows, spend tracking, brand kit consistency, and automated channel monitoring via webhook or polling.

unbridled bridges agent messaging to facebook messenger, whatsapp, instagram, linkedin, twitter, signal, telegram, and discord via a user's beeper account. sends and reads messages across all networks with end-to-end encryption and cross-signing bootstrapped from recovery keys.

backup-to-r2 configures encrypted off-site backups of openclaw installations to cloudflare r2 using restic, with portable restore capability and optional task scheduler automation for daily snapshots.

brand-impersonation-response provides incident playbook for deepfakes, fake apps, spoofed domains, and credential harvesting scams. covers verification, platform-specific takedown sequencing, multi-audience comms, and post-incident hardening with internal protocols.

guardrail skill for customer support replies that enforces fact verification, policy compliance, private data minimization, and prevents unauthorized promises. provides decision framework and pre-flight checks for support agents.

skill-reliability-design guides quality managers through fmea analysis, reliability metric calculation, and design verification during product development. covers principles, quantitative assessment, and continuous improvement mechanisms.

project-loop manages multi-step work that survives interruption through file-backed state, approval gates, and bounded retry. executes self-clearable tasks continuously, pauses at real blockers or human decisions, and resumes from verified checkpoints.

agentmfa gates sensitive agent actions (deployments, deletions, transactions) behind human approval via push notifications and biometric verification. agents explicitly invoke approval workflows through mcp tools, with blocking waits and rejection handling.

minio-aiops provides 31 governed tools for object storage diagnostics and safe operations: capacity root-cause analysis, bucket exposure auditing, lifecycle gap detection, and healing health checks, wrapped in a bundled audit and undo harness with encrypted credential storage.

kabuzz-marketplace provides 49 tools across buying, selling, negotiating, and order management for ai agents. covers search, listing creation with ai-generated metadata, offer negotiation, and shipping. integrates payment methods and configurable spending controls.

mechanical-keyboard-buying-consultant interviews users on use case, environment, typing feel, layout, connectivity, and build priorities to deliver a structured, brand-neutral spec recommendation covering switch type, actuation force, form factor, keycap material, and regional layout standards.

posteahora exposes a cli for scheduling and publishing social posts across nine platforms. covers authentication, account discovery, media upload, post creation with scheduling, and analytics. requires explicit account ids and iso 8601 timestamps.

newsletter-operator covers end-to-end newsletter workflows including research, drafting, review, promotion, growth, and monetization. it routes specialized work to focused skills and enforces approval gates before destructive or external actions.

seven-questions-test operationalizes peter thiel's framework for stress-testing business ideas before commitment. it forces written, evidence-backed answers to seven questions (engineering, timing, monopoly, people, distribution, durability, secret) and gates advancement on scoring anchors, not vibes. includes adaptive coaching modes for novices and calibration against real outcomes (solyndra vs tesla).

fetch-evm-logs pulls and parses event logs from any evm smart contract across all evm-compatible chains. requires chainid, contract address, and abi (auto-discovered or user-provided), then scaffolds a typescript project and outputs structured logs.

onthecheap-mcp queries free and cheap event listings across 14 us cities with daily calendars, searchable posts, and category filters. read-only access requires no auth. always specify the city.

runapi-grok exposes grok-4.5 via openai-compatible sdk with streaming, function tools, and structured json output. covers setup, chat completions, reasoning effort levels, and multi-protocol availability.

homebutler covers homelab server operations including status checks, docker management, multi-server ssh, backup verification, and self-hosted app deployment via cli with mcp server support for ai integration.

perfguard scans codebases for 40+ performance anti-patterns across python, javascript, ruby, and java, detecting n+1 queries, sync i/o, missing pagination, and memory leaks with local regex-based analysis and tiered reporting via git hooks and audit reports.

human-test recruits real humans to test products and delivers structured usability reports with nps scores, video analysis, and ai-generated code fix suggestions. supports both hosted and self-hosted deployment.

easydoc-extract wraps the easylink async document extraction api, supporting four modes (universal, flash, business license, org-code) with optional schema-driven or prompt-based field extraction. covers submit-poll-normalize lifecycle with file validation and output normalization.

storage-cleanup reclaims disk space across macos and linux by targeting 12+ cleanup categories including caches, trash, old kernels, and container artifacts. dry-run mode previews changes safely before execution.

mcp-server-setup-kit provides a structured five-phase protocol for connecting claude to github, notion, slack, and linear via mcp servers. includes pre-built config templates, verification prompts, and a decision tree for common setup failures.

aeon is a scikit-learn compatible time series ml toolkit covering classification, regression, clustering, forecasting, anomaly detection, segmentation, and similarity search with working code examples and algorithm selection guidance.

computer-task-execution prioritizes reliability and verifiability over cleverness, routing tasks through apis and browser execution before falling back to ui automation, with mandatory post-action verification and learned pattern reuse per software target.

save persists agent conversations to a three-tier local memory system: human-readable log, fts5-indexed search database, and full markdown files. no external dependencies, python stdlib only.

demandex mines reddit complaint posts across 70 communities to surface product demand signals, returning scored opportunity cards with evidence and permalinks. pays per-call via x402 usdc on base, with free endpoints for schema discovery and no mandatory accounts.

psychotherapy-case-formulation guides licensed clinicians and supervisees through structured biopsychosocial intake synthesis, producing a draft 4Ps formulation with provisional DSM-5-TR diagnosis, risk flags, and questions for licensed review before any clinical use.

phy-post-forensics extracts 12 linguistic and structural features from batched social posts, groups by engagement tier, and surfaces which content patterns correlate with your highest-performing content. outputs a data-backed content blueprint rather than raw metrics.

improve-hn-karma generates daily batches of 4-5 tailored comments on rising hacker news threads, identifying underexplored technical angles and delivering copy-paste-ready posts with timing guidance to build karma on low-reputation accounts.

economic-moat evaluates whether a company has structural, durable competitive advantage by testing five sources (intangible assets, switching costs, network effects, cost advantages, efficient scale) against the greenwald mechanism test, trajectory analysis, and smart-attacker simulation.

prompt-debugger provides systematic diagnosis of failing prompts through structural analysis, failure mode classification, and iterative rewrites. includes reference patterns, anti-pattern detection, and quality scoring.

linkfox-echotik-product-video retrieves tiktok shop promotional videos for a given product, surfacing engagement metrics, estimated sales attribution, and creator ids to help sellers identify high-performing content and influencer patterns.

agent-skills-by-notifyer covers account lifecycle management, team administration, workspace configuration, and whatsapp connection validation via console api. prerequisite checks and explicit plan-gating rules prevent common integration failures.

runware-run codifies the execution contract for calling runware apis across modalities. it enforces schema validation before requests, distinguishes sync (images) from async (video/audio/3d) execution, and provides cost-gating via dry-run headers.