my-humanizer rewrites ai-generated text into natural prose, measures transformation locally, and optionally verifies results via structured metrics without exposing content.

kosmi-dj automates video queueing and playback in Kosmi watch party rooms via agent-browser. it connects to rooms, plays videos by url, and supports looping dj mode with persistent sessions.

flyelep-image-clarity-enhance wraps a single-purpose http api for upscaling image sharpness at three configurable strength levels. supports single and batch processing via comma-delimited url strings, with explicit format and dimension constraints.

claude-code-cli delegates coding tasks to Claude Code CLI in background or interactive mode, supporting PTY for confirmations, headless pipe mode for automation, session continuity, and granular tool/permission controls. covers feature building, pr review, refactoring, and parallel issue fixing.

ai-chatbot-prompt-builder covers four modular prompts for generating system instructions, personas, safety guardrails, and faq datasets. targets businesses deploying custom llm assistants without prompt engineering expertise.

academic-figures generates publication-quality charts (bar, heatmap, scatter, line, box, forest, violin) from json/csv data with native cjk support, journal color palettes, and dual png/svg output. local execution keeps data private.

finlab covers quantitative trading strategy development via a 5-step workflow: data fetching, factor creation, position construction, backtesting, and optional live order execution. includes preset data catalog, dataframe methods, risk parameters, and lookahead bias prevention.

pdf-translate extracts and translates pdf documents to chinese markdown, then renders typography-aware pdfs via weasyprint. handles structural hierarchy, code blocks, tables, and cjk fonts with explicit quality rules.

wangjipeng-meeting-transcript-to-summary extracts structured summaries from meeting transcripts, organizing decisions, action items, and open questions with explicit speaker and timestamp handling.

molty.pics enables openclaw bots to create and share ai-generated images, comment, like, and follow peers. the platform enforces rate limits and requires human claim activation before posting.

claude-agent-team-workflows provides a reusable orchestration framework for 5-agent teams across domains via role remapping and four control-flow patterns (sequential, parallel-merge, iterative-review, fan-out-fan-in), with strict 7-step coordination protocol and lead discipline rules.

generates social media cover images by templating html based on user-supplied titles and platform dimensions, then converts to png via headless browser. supports 11 major platforms with preset aspect ratios and color scheme suggestions.

market-research-pro provides a seven-step framework for generating structured market reports including tam/sam/som sizing, porter's five forces analysis, pestel scanning, and competitive landscape mapping. outputs are markdown reports with explicit source citation and assumption flagging.

deso-research queries farcaster, lens, nostr, and bluesky via the deso-ag cli to search topics, extract trending terms, and analyze engagement across decentralized social networks. outputs are formatted for agent consumption with built-in analysis guidelines.

clean-json-toolkit provides local inspection, querying, flattening, and validation of json and jsonl files using pure python standard library, with jq-style path syntax and reversible flatten/unflatten operations.

antv-g6-graph covers g6 v5 initialization, layout, styling, and interaction patterns for network, tree, and flow diagrams. heavy emphasis on v4 migration pitfalls and api correctness constraints.

bookforge-bullseye-channel-selection applies a five-step framework to systematically evaluate 19 customer acquisition channels and isolate the three most promising for parallel testing. guides founders away from channel bias through exhaustive brainstorming and ranked tiers.

agent-browser-0-2-0 automates web interactions via CLI commands, offering snapshot analysis, element interaction by reference, and session state persistence for form filling, data extraction, and ui testing.

hopeids provides inference-based intrusion detection for ai agents with configurable per-agent risk thresholds, pattern matching against 108 detection patterns, quarantine storage (metadata only), and human-in-the-loop telegram alerts. blocks or warns on jailbreaks, prompt injection, credential theft, and social engineering attempts.

sn94-competition-poster guides creation and packaging of BitSota competition proposals and replayable task repositories, enforcing narrow patch surfaces and backend-agnostic validator design.

skillfactory codifies a four-phase sop (capture, collect, distill, publish) for turning raw ideas into versioned, distributable openclaw skills. includes structured upgrade workflow with semantic versioning and clawhub publishing automation.

red-team orchestrates multiple ai agents with opposing viewpoints to debate decisions through structured rounds, producing conviction scores and synthesis briefs for stress-testing ideas and validating strategy.

generator-buying-consultant guides users through structured load-sizing calculation (running watts, surge watts, safety margin) to identify required generator rated output, voltage, fuel type, and features independent of brand bias, using appliance reference tables and regional voltage standards.

performance-optimization covers frontend and backend techniques to reduce load times, bundle size, and database latency. covers react memoization, code splitting, image optimization, query tuning, and caching patterns with measurement frameworks.

headteacher workbench bootstraps ai-native class management on feishu base, notion, or obsidian. guides backend selection, schema installation, student record operations, and generates office artifacts (docx, xlsx, pptx) from structured data.

feishu-backup downloads files from feishu group chats to a local directory with configurable filters for name prefix, keyword, file type, and time range. execution depends on a shell script wrapper that validates file integrity.

pessimistic-offline-lock-implementer guides end-to-end implementation of pessimistic offline locking for multi-request business transactions. covers lock type selection, durable lock manager design, timeout and release mechanics, aggregate-level locking, and implicit lock wiring to prevent gaps.

phenoskill extracts clinical phenotypes and medication entities from user-provided text using the phenosnap library, with built-in pii redaction and local-only processing. outputs timestamped json with detected entities.

sentinal-redis diagnoses redis server health, memory usage, slow queries, and bullmq queue state through read-only redis-cli commands. covers connectivity, memory fragmentation, job failure inspection, and stale worker detection with decision trees for common failure modes.

image-with-comfyui wraps a local comfyui server for text-to-image, image-to-image edits, and image-to-video generation. supports z-image, sd3.5 medium, qwen image edit, and wan2.2 models with automatic prompt formatting and vram cleanup.

google-docs-cli wraps porteden to manage google docs lifecycle: creation, content reads/edits, sharing, permissions, and deletion via cli. supports markdown import, bulk ops from json, and ai-optimized output flags.

paragraph enables automated web3-native blogging via paragraph.com with onchain post storage, subscriber management, tokenized engagement through social coins, and feed discovery. covers 19 distinct api tools for content publishing, audience building, and token economics.

monthly audit that inventories agent stacks for zombie crons, unused api keys, dead scripts, and stale subscriptions, classifying findings by delete/replace/upgrade/healthy and outputting a ranked cleanup brief.

fetches economic calendar events for a date range (default 7 days) via tradingeconomics api or yahoo finance fallback, with optional cron scheduling for daily briefings.

cfshare wraps cloudflare quick tunnel to expose local ports and files as temporary public urls via cli, with structured json output and audit trail support.

higgsfield-soul-id trains a reusable face model for identity-faithful image and video generation. one-time setup returns a reference id for use across soul-powered generation commands.

persistent-memory adds a three-layer (markdown, vector, graph) recall system to openclaw workspaces, enabling agents to maintain semantic understanding of decisions and institutional knowledge across sessions with automatic directive indexing.

llama-params-optimizer provides a systematic 4-phase 10-step methodology for tuning local llm startup parameters on any hardware. core strategy centers on finding the gpu memory sweet spot where context window maximizes without degrading throughput, validated through control variable testing across batch size, threads, quantization, and flashattention settings.

cpbox-answers provides openai-compatible ai-grounded answers via web search with two modes: single-search for speed and research mode for deep multi-iteration synthesis. supports streaming, citations, and x402 payment.

solo-landing-gen produces landing page copy (hero, features, a/b headlines, cta, seo tags) from project prd. scaffolds astro pages when detected. requires product context to avoid generic output.

skill downloads recent or specific videos from public tiktok accounts via yt-dlp, supporting single/bulk/metadata-only modes with username normalization and error recovery for rate limits and access restrictions.

personal-api scaffolds an ai-readable identity layer in obsidian vaults by installing me.md and agent.md contracts, with optional knowledge palace v2 structure. routes users through setup, customization, knowledge production, and validation workflows.

seo-keyword-research-aisa-api combines site crawling with dataforseo api calls and llm reasoning to identify keyword opportunities, cluster by intent, and recommend content pages. built for saas and product sites with competitor and serp analysis.

skill-vetter-bwm provides a structured checklist for security vetting of ai agent skills before installation, covering source validation, code review red flags, permission scope analysis, and risk classification with decision output.

spotify-playlist-curator chains multiple spotify and reccobeats apis to create, refine, and discover playlists from seed tracks, listening history, genre profiles, and audio feature targets. covers taste persistence, eight api regressions, and tool selection heuristics.

llm-trading-agent-security covers layered defense patterns for transaction-signing agents: prompt injection detection, spend limits, pre-send simulation, circuit breakers, and wallet isolation. includes code examples and a pre-deploy checklist.

biz-in-a-box provides a protocol and file structure for agent-native double-entry bookkeeping across any entity type, with append-only journaling, validation, and report derivation capabilities.