tiktok-script generates production-ready short-form video scripts optimized for tiktok's discovery engine, built around a brutal 1-second hook, watch-time retention, and native (non-ad) tone. outputs a shootable beat-by-beat doc plus veo prompts for faceless content.

wechat-publisher-md2wechat-draftget orchestrates official weChat draft creation from markdown or html with mandatory backend verification via draft/get, enforcing clear success criteria that distinguish submission receipts from real backend presence.

ai-qa-agent enforces quality gates across code correctness, data integrity, brand voice compliance, and document formatting. executes five-phase review pipeline producing severity-ranked findings and clear pass/fail verdicts with actionable remediation.

kling-3-0 wraps all six kling 3.0 endpoints (standard, pro, 4k across text-to-video and image-to-video modes) via runcomfy cli. covers tier selection logic, multi-shot prompting patterns, audio generation, and endpoint invocation with concrete bash examples.

delivery-proof coordinates exact-approved delivery actions, validates against pre-defined acceptance criteria, and produces evidence-labelled records with explicit privacy safeguards, failure-mode awareness, and separation of human from agent work.

download-dem orchestrates deterministic multi-source digital elevation model discovery, download, resumption, mosaicking, and validation across microsoft planetary computer, aws open data, opentopography, usgs 3dep, and nasa earthdata. supports city-to-country aois with exact vector masking, resumable transfers, and provenance reporting.

scavio-google-maps wraps three google maps endpoints (search, place details, reviews) into structured json. covers localization via lat/lng or country code, pagination, and multiple identifier formats across endpoints.

competitor-show-tracker ranks upcoming trade shows by competitor density using lensmor exhibitor records. accepts multiple competitor names, aggregates events, filters to future dates, and surfaces shows where competitive clustering is highest for strategic booth investment decisions.

amazon-sp-api wraps seller central operations (orders, inventory, reports, finances) behind a python script that handles lwa refresh and sigv4 signing internally, exposing a clean json interface across all amazon marketplaces and regions.

coverage-gap-analysis maps organisational risks against insurance policies to expose gaps, thin coverage, and accidental overlaps. outputs a risk-by-coverage matrix, deductible stack review, and severity-ranked recommendations for closure.

bug-diagnosis guides systematic root-cause isolation through reproduction, narrowing, hypothesis ranking, and verification. produces a minimal repro, pinpointed location, ranked test plan, and regression guard against re-occurrence.

board-deck-narrative structures quarterly and special board presentations with explicit slide-by-slide guidance, narrative principles, and quality checks. it anchors on honest framing, specific asks, and discussion-first design rather than presentation volume.

cohort-analysis structures retention, ltv, and behavioural analysis with explicit input validation, step-by-step procedure, mutually exclusive segment definitions, and anchored recommendations. output templates enforce grounding in observed data and flag unsupported claims.

adversarial-design-review runs a panel of specialist agents against risky system designs (drivers, formats, concurrency) before implementation to surface concrete failure modes and fixes, validating load-bearing choices and deferring expensive debugging to the design phase.

axis-trust enables verification of ai agent identity, behavioral reputation (t-score 0-1000), and economic reliability (c-score aaa-d) via cryptographic auids. lookup any agent's trustworthiness before delegation, transaction, or data sharing.

personal-board-of-directors convenes five archetypes (operator, skeptic, cfo, coach, customer) to pressure-test a decision through distinct memos, disagreement mapping, and a vote with stated decision rules and flip conditions.

early-access-designer structures product launches through a waitlist-to-ga stage ladder with quantified graduation criteria, cohort gating rules, tester feedback loops, and referral mechanics. outputs are submitted to launch-registry for canonical record.

overwatch enforces mandatory communication during agent execution, requiring status updates at task start, progress milestones, sub-agent spawns, errors, and completion. prevents silent failures and keeps humans informed of work in progress.

container-host-aiops provides governed single-host Docker and Portainer operations through 34 MCP tools, bundling audit, policy, budget guards, and undo recording. covers container/image/volume/network inspection, three flagship analyses (restart-loop RCA, resource pressure, bloat), and guarded writes with dry-run support.

agent-peer-lan establishes direct collaboration between two openclaw agents on a trusted local network subnet, eliminating vpn dependency and third-party relay overhead while keeping all traffic and credentials local.

3pl-evaluator guides sellers through systematic cost modeling and vendor comparison using normalized landed-cost math, weighted scoring, and reference validation to avoid hidden fees and mismatched partnerships.

dark-social-attributor decomposes direct traffic into plausible dark-social share flows, specs utms and self-reported attribution fields, and labels every estimate as proxy or heuristic. it stops at organic owned loops and avoids paid-platform reconciliation, closed-platform automation, and conversion claims.

workspace-git-setup initializes git repositories with a hardened .gitignore, detects large files and tracked secrets, and offers audit and dry-run modes. pure bash implementation, no external dependencies beyond git.

openclaw-thumbnail-forge generates professional video thumbnails locally using frame ranking heuristics, text composition, and platform-specific export. no ai apis, no remote calls, pure ffmpeg and pillow.

carsxe wraps 12 vehicle data endpoints covering vin/plate decoding, market value, history, recalls, liens, obd diagnostics, and image ocr. request routing is explicit and chaining strategies are documented with working examples.

power-law-distribution guides resource allocation and risk modeling in domains where outliers dominate totals. it surfaces systematic errors in gaussian-based planning and replaces average-case thinking with tail-aware strategy.

exa-search-rust wraps the exa ai neural search engine as a native rust binary, covering web search, similar-page discovery, and url content fetching with explicit mode selection guidance and performance benchmarks.

robotomail provides api-first email infrastructure for autonomous agents, covering send, receive, threading, webhooks, sse streaming, and custom domain setup with explicit rate limits and attachment handling across multiple payload shapes.

message-system-architect authors the durable brand message hierarchy (main narrative, three value pillars traceable to positioning themes, per-persona proof points, tagline) that seeds the narrative-registry canon and upstream of per-launch message houses. it enforces internal consistency and claim provenance labeling (measured, user-provided, needs-source).

zipcracker bundles zip recovery and cracking workflows for ctf and authorized security work, supporting dictionary, mask, crc32, known-plaintext, and bkcrack attacks with built-in pseudo-encryption repair and template matching.

north-star-metric guides teams to pick a single leading indicator of customer value that predicts revenue. covers activation triggers, a six-step audit process, stress-testing for gaming, and quarterly re-evaluation across product stages.

woocommerce-agent-guardrails enforces tiered approval gates and hard caps on woocommerce writes, with mandatory dry-run previews before tier 2-3 actions, circuit breakers on repeated failures, and defense against prompt injection in imported data.

fetches complete product detail for a single ozon sku, returning sales metrics, revenue, stock, category rank, daily trends, seller and brand info. designed for single-product analysis, competitor teardown and listing diagnostics.

launch-feedback-synthesizer clusters feedback from launches into themed digests with frequency and severity, specs a status loop with notifications and duplicate merging, produces shipped-change announcements, and designs compliant review-harvest protocols without executing outreach or amplification.

landing-experience-checker preflights destination pages against ad copy for message continuity, above-the-fold offer visibility, load speed, form friction, and mobile rendering, producing a pass/fix punch list before campaign launch.

wjs-reframing-video crops and reframes video between horizontal and vertical orientations by tracking the active speaker's mouth motion via MediaPipe face landmarks, outputting a crop plan sidecar and the reframed video.

converts figma and pencil designs into layered react code via feature-sliced design and tailwind v4, with explicit rules for tokens, accessibility, server/client boundaries, and repo-specific constraints.

review-problem fetches unevaluated research problems from an mcp server, surveys related literature, scores each on 5 value and 5 difficulty metrics (1-5 with rationale and evidence), then submits via post_problem_evaluation to place it in a value-difficulty quadrant.

alphagbm-company-profile builds persistent research profiles for tracked stocks, auto-populating fundamentals, valuation bands, red flags, and event timelines. supports create, list, refresh, and delete operations across tiered subscription limits.

clawkey binds openclaw agents to verified human owners via ed25519 key signing and veryai palm verification. agents initiate registration by posting a signed challenge, humans complete oauth verification, and third parties can verify agent ownership.

k-factor-viral decomposes viral growth into invite rate (i) and conversion rate (c) to calculate K = i * c, distinguishing actual exponential growth (K > 1) from vanity metrics. covers measurement discipline, wave decay, and separate optimization levers.

defluff detects ai-authored, ai-assisted, or human-written emails; reverses suspected ai prompts; classifies urgency and intent; and flags common scams (invoice fraud, bec, phishing, fake recruiters, crypto pitches). handles single messages, threads, and batches with no conversational padding.

spraay-compute rents gpu time and runs ai model inference (llm, image, video, speech) via x402 micropayments in usdc, plus optional prepaid compute-futures accounts with tiered discounts up to 15 percent.

sixtyfour-skill wraps a live-web people and company research api, enabling lead enrichment, email/phone discovery, lead scoring via custom criteria, and batch workflows. returns confidence-scored structured data with source attribution.

deliberate-practice targets skill plateaus by isolating representational gaps and structuring feedback-rich repetition under cognitive load, grounded in ericsson's deliberate practice framework. distinguishes itself from mere execution and automaticity maintenance.

golang-spf13-cobra covers command tree design, hook chains, flag parsing, args validation, shell completions, and testing patterns for go cli applications. emphasizes runE over run, persistent flag inheritance, and testability via output redirection.