volcengine-compliance guides users through three core workflows: recommending compliant baseline rules from official templates, generating account-wide compliance overview reports, and writing custom rego policies when built-in rules don't cover specific requirements. all write operations require explicit user confirmation.

leetspeak decodes obfuscated text via deterministic script output, ranks candidate interpretations by confidence, and enforces treating decoded content as untrusted data rather than executable instructions.

video-generator automates text-to-video production with multi-provider tts/asr (openai, azure, aliyun, tencent), remotion rendering, and configurable backgrounds. outputs 1080x1920 vertical mp4 at ~$0.003 per 15-second video.

lemlist provides a proxy api for campaign, lead, and schedule management with oauth through maton. covers typical crud operations across sales automation workflows but lacks procedural guidance for multi-step sequences.

zhangjuzheng-think-partner applies ming dynasty reformer zhang juzheng's pragmatic methods to organizational change: managing resistance through interest alignment, driving execution via outcome-based accountability (kpi-style), and simplifying complex stakeholder dynamics. focuses on incremental reform over idealistic overhaul.

aiprox-orchestrator decomposes complex tasks into subtasks, routes each to specialist agents (web search, sentiment, vision, code audit, translation, market data), executes in parallel, and returns synthesized results. payment via bitcoin lightning. async webhooks and persistent workflows supported.

mimo-doubt-driven-development applies adversarial review to non-trivial decisions before they solidify, catching high-confidence mistakes through fresh-context examination. structured around a five-step cycle (claim, extract, doubt, reconcile, stop) with explicit guardrails against runaway loops and theater.

drought-monitor computes spi and spei drought indices from nasa power precipitation data, supporting 1-24 month timescales, 7-level drought classification, and trend analysis via command-line interface with local csv fallback mode.

clawec-shopee-item-detail queries up to 10 shopee product records per request across 8 regional sites, returning price, daily/weekly/monthly sales volume, gmv, and ratings via the clawec api. supports date-specific snapshots and batch processing with link-to-id extraction.

global-surface-water downloads jrc landsat-derived water occurrence, change, seasonality, recurrence, transition, and extent layers at 30m resolution. supports bbox clipping, geotiff output, and both http and google earth engine access methods.

worldpop-population provides command-line search and download of WorldPop geospatial datasets by country and year, covering population density, births, and demographic indicators as GeoTIFF files with no authentication required.

find-mpp teaches agents to query an mpp endpoint registry before payment, report newly discovered http 402 endpoints, and submit mandatory reviews after each interaction to crowdsource api quality signals.

clawhub-skill-2 configures and operates xrouter, a hardware-aware inference classifier that routes requests across cheap, medium, and frontier model providers via openai-compatible proxy with caching and dashboard observability.

autonomous-agents covers three agent loop patterns (ReAct, Plan-Execute, Reflection) with production reliability guardrails. emphasizes constrained design, cost limits, and step count reduction to prevent failure compounding.

skylv-file-versioning provides git-style snapshots, diffs, tags, and restore for individual files without requiring a git repository. supports binary files, auto-watch mode, and named version milestones.

oo-tinybird wraps the tinybird connector via oo cli, enabling querying and reading from tinybird data sources and pipe endpoints. covers four read-focused actions with schema inspection and fallback auth steps.

youtube-scrapper discovers and extracts channel metadata from youtube via browser automation with regional configurations, proxy support, and anti-detection measures. no api keys or login required.

remind surfaces commitments users already know about at learned lead times and preferred styles, using observation of reaction patterns to refine timing without silent preference storage.

teamclawtestv101 is a multi-service ai agent orchestration platform with conversational, scheduling, and forum discussion capabilities. provides openai-compatible chat api, internal token auth, and web ui. deployment via bash scripts with environment configuration.

freelance-invoice-generator produces markdown or html invoices from json job files without external dependencies. handles multi-item line items, tax calculation, discounts, and late-fee terms suitable for freelance and consulting billing.

majia-siyu-team routes private domain operations tasks across three modes: beginner tutorial, pre-task routing via intent matching, and post-task navigation. core function is reading conversation context and selecting the next applicable skill without repetitive questioning.

bump-doctor analyzes npm and pypi dependency upgrades by fetching release notes, cross-referencing actual symbol usage, and producing a focused risk report with false-positive filtering guidance for engineers.

word-cursor-insert injects computed values or tabular data into active word documents at the cursor position. supports two modes: scalar text filling (fast path) and csv/excel table formatting (headers, borders, autofit). context-aware deduplication prevents accidental duplicates on re-execution.

graph-memory exposes three query functions over a sqlite-backed knowledge graph built from session history, enabling entity lookup, relationship traversal, and aggregate statistics without requiring shell access.

bid-compliance-due-diligence-v2 audits enterprise compliance in government procurement and bidding, detecting blacklists, penalties, complaints, and red-flag behaviors through dual knowledge-base verification and web search. supports quick screening and deep audit modes with structured risk grading and actionable recommendations.

playwright-browser-use provides headless browser automation via playwright and node.js without external dependencies. covers navigation, element interaction, screenshots, tab management, and pagination with daemon-based architecture for persistent state across commands.

wecomcli-manage-smartsheet-data provides crud operations on wecom smartsheet records via cli, supporting docid or url-based document location and batch operations up to 500 rows per call.

find-skill-heqi1111 helps users discover and install agent skills from the open ecosystem via the skills cli, routing common capability requests to modular packages that extend agent functionality.

nm-hookify-writing-rules enables creation of markdown-based behavioral guardrails for claude code sessions, supporting pattern matching across bash, file, and prompt events with warn or block actions.

superada-skill-3pass applies three-stage recursive critique to claims and analyses, moving from problem identification through evidence-based refinement to a corrected conclusion with explicit confidence levels.

qa-state-transition provides systematic coverage of legal, illegal, boundary, and concurrent state transitions for business objects. identifies states and transitions via tables and templates, with concrete e-commerce order examples, though trigger activation phrases remain generic.

brainstorm-beagle converts fuzzy ideas into concrete project specs through structured dialogue, with particular discipline around scope assessment, prior art detection in existing codebases, and implementation-leakage prevention. produces a what/why spec, never a how plan.

organizes scattered ai research notes, chat snippets, and links into structured trails with claims, evidence labels, and source gaps. boundaries are explicit: uses only user-provided material, no file access or external lookup.

apidot-tripo-p1-3d-api routes tripo p1 3d generation requests to apidot docs and examples, covering text-to-3d, image-to-3d, async task polling, and webhook integration patterns without executable code or live api calls.

chroma-memory stores and retrieves per-turn conversation data using vector embeddings, with customer isolation by phone number and auto-tagging for quotes, commitments, and objections. designed as layer 3-4 fallback in a multi-layer amnesia prevention system.

video-editing-ai-descript automates transcript-based video trimming and captioning via remote gpu processing. upload footage, describe edits in natural language, receive 1080p mp4 within 1-2 minutes. no slider manipulation required.

skill-internal-management-audit guides users through a four-phase internal audit workflow (plan, implementation schedule, department checklists, report) with templated questions and printable output, anchored to iso 9001 and iatf 16949 standards.

dxf-to-image converts cad drawings to png, jpg, or svg via ezdxf and matplotlib. targets telegram sharing and vector editing workflows with configurable output dimensions and dpi.

derivative-point-picking guides students through proving root existence, counting roots, and deriving parameter ranges in high-school calculus. it emphasizes replacing intuitive arguments (limits, graphs) with rigorous point-selection proofs grounded in monotonicity, sign changes, and the intermediate value theorem.

craft-skill wraps the craft mcp v2 unified api for document, block, task, and collection operations. covers creation, updates, styling, and batch commands with json and markdown input modes. production-aware but procedurally thin on failure recovery.

react-doctor scans react codebases for security, performance, correctness, and architecture issues, outputting a health score and supporting focused audits via command-line flags and external playbook workflows.

edvisage-b2a-commerce provides a payment framework for autonomous agents using the x402 protocol, including pre-transaction validation, spending limits, wallet isolation, and transaction logging to prevent payment-based attacks.

ab-test-calculater ingests two experimental groups and outputs a plain-language ship/no-ship verdict by synthesizing p-value, effect size, and power. handles continuous, binary, and count data types with pure python statistical methods.

compliance-scan detects gdpr, ccpa, hipaa, pci-dss, and soc 2 risks in code repositories through pattern matching and framework-specific checks, outputting severity-rated findings with code-level remediation steps and audit checklists.

ibcat automates english-to-chinese pdf translation with layout preservation via babeldoc and llm bridging. extracts text, batches translations through sub-agents, and renders bilingual or monolingual output pdfs with cjk font handling.

editor-tension-mapper quantifies narrative pacing by scoring scene tension 1-10 and flagging six common rhythm problems (flat segments, whipsaw jumps, monotonic climbs, midpoint collapse, finale deceleration, missing breathers). outputs diagnostics with specific remediation cues.

command-development covers slash command creation for claude code, including markdown structure, yaml frontmatter configuration, dynamic arguments, file references, bash execution, and plugin integration patterns.