video-edit routes user requests to appropriate runcomfy models (wan 2.7, kling motion-control, lucy restyle) based on detected intent for video transformation tasks.

content-modeling-best-practices outlines foundational principles for structured content design in headless cms environments, emphasizing reusability and maintainability, but lacks concrete implementation steps and decision-making frameworks.

cloudbase-guidelines outlines tencent cloudbase development practices across web, mini programs, and native apps, covering authentication, databases, functions, and hosting with mcp tooling recommendations.

dt-obs-logs enables querying and filtering dynatrace log data via dql for troubleshooting. covers severity filtering, pattern matching, error rate calculation, and aggregation by dimension.

ltx2 generates short video clips from text or images using the lTX-2.3 model on Modal infrastructure. supports resolution/frame tuning and quality modes but lacks error handling guidance.

research-engineer enforces scientific rigor and rejects invalid problem constraints. prioritizes theoretical correctness over convenience, with explicit protocols against hallucination and oversimplification. incomplete procedural coverage limits real-world applicability.

isms-audit-expert provides iso 27001 audit frameworks and control assessment taxonomies, but lacks actionable procedural depth, concrete trigger conditions, and failure mode handling. reads as a knowledge base outline rather than an executable skill.

simulates a five-reviewer academic peer review process with distinct personas (eic, three domain reviewers, devil's advocate) to produce editorial decisions and revision roadmaps. field detection and multi-perspective assessment included.

figma-generate-library orchestrates multi-phase design system builds in figma by coordinating 20-100+ api calls. prerequisite skill figma-use must load. enforces quality patterns from production systems like material 3 and polaris.

attach-review-to-pr covers adding line-specific comments to pull requests via github cli api or mcp tools, with fallback paths for unavailable endpoints and guidance on single vs batch operations.

test-data-generation covers factory and builder patterns for creating realistic, reusable test fixtures. includes code examples and best practices but lacks decision logic and failure recovery.

persona-hr-coordinator automates onboarding workflows via google workspace (calendar, drive, chat, gmail) with pii sanitization support, but lacks explicit decision logic and failure handling.

llm-prompt-injection covers direct and indirect injection attack vectors including rag poisoning, tool abuse, and encoding bypasses, but lacks actionable procedures and concrete trigger conditions.

android-aso provides comparative context for google play optimization versus ios, covering indexing differences, metadata character limits, and platform-specific constraints. lacks actionable steps.

asc-localize-metadata automates app store metadata translation by pulling english or source locale content, processing it through llm translation, and syncing results back to app store connect via asc cli.

longbridge-technical documents trigger conditions for multiple technical analysis frameworks across asian and us equity markets, but lacks procedural depth, actionable steps, and failure mode handling.

ansible-automation provides a framework for infrastructure provisioning and configuration management via playbooks and roles, with a do/dont checklist but minimal procedural depth or failure recovery guidance.

industry-research outlines a methodology for multi-source data gathering and analysis across sectors, but lacks concrete trigger conditions and decision logic. execution relies heavily on manual interpretation of frameworks rather than automated workflows.

microsoft-agent-framework covers language detection and routing between .net and python implementations, with heavy reliance on external documentation for actual implementation guidance.

flash provides a local development-to-cloud deployment workflow for runpod serverless workloads, bridging dev testing at localhost:8888 with remote gpu/cpu provisioning via cli commands.

verified-agent-identity enables decentralized agent identity creation on ethereum via ERC-8004 standard, linking agents to human owners, and handling 402 payment authentication through signed headers. local identity storage with optional encryption.

figma-integration provides guidelines for mapping design artifacts to code via mcp servers and manual extraction workflows, covering color, typography, layout, and asset handling without actionable decision logic or failure recovery.

shared-world organizes canonical fiction data in wiki format for multi-author consistency. lacks concrete procedures for contradiction detection, version control, or contributor workflows despite framing potential.

illustrator bridges agent shell commands to adobe illustrator via flue, a local extendscript transpiler. enables bounded document inspection and editing without mcp server.

agentica-infrastructure documents 11 multi-agent coordination patterns and their apis, with constructor signatures and class methods. lacks procedural depth, edge case handling, and trigger specificity for practical deployment.

english-learner passively intercepts and corrects user prose across english, chinese, and seven other languages using tiered output modes and persistent vocab logging, but lacks explicit failure handling and complete procedural detail.

wireframe generates UI mockups and layouts in drawio XML format with platform-specific component libraries for web, iOS, and Android. delegates most structural decisions to a parent drawio skill and provides platform dimensions and spacing guidelines.

android-development provides reference guidance on google architecture patterns and nowInAndroid conventions, but lacks executable procedures. structured as a navigation layer to external documents rather than actionable skill content.

improve-codebase-architecture identifies architectural friction points and refactoring opportunities to deepen shallow modules, guided by domain language consistency and decision records. targets testability and codebase navigability.

seo-keyword-strategist analyzes keyword density, identifies entities and lsi variations, and flags over-optimization in provided content. outputs a structured keyword strategy package with placement recommendations and search intent assessment.

pentest-commands provides a flat command reference for common penetration testing tools (nmap, metasploit, hydra, sqlmap) organized by tool category. lacks procedural flow, decision logic, and actionable workflows for realistic assessment scenarios.

rust-engineer covers idiomatic rust authoring focused on memory safety, ownership patterns, and trait design. provides a five-phase workflow for analysis, design, implementation, error handling, and validation. lacks specific trigger phrases and actionable decision trees.

aflpp enables multi-core fuzzing campaigns for c/c++ codebases using diverse mutation strategies, positioned as a mature alternative to libfuzzer for throughput-constrained projects.

rclone provides a setup check and partial procedure for file transfers across cloud providers, but lacks complete operation workflows, clear decision logic, and robust error handling documentation.

extension-object-storage enables off-chain file/object storage with on-chain references via caffeine ai, requiring coordinated backend and frontend package installation and explicit storage.externalblob typing for all file fields.

extension-posting-to-x documents x-client mops integration for caffeine builds posting to twitter, emphasizing mandatory oauth 2.0 pkce compliance and architectural guardrails against raw http anti-patterns.

using-ui-bundle-salesforce-data specifies mandatory use of the salesforce data sdk for record operations within ui bundle contexts, covering graphql queries and rest fallback patterns.

brand-designer provides code templates and asset organization for logos, color systems, and typography, but lacks actionable intent definition, specific trigger phrases, and failure recovery guidance.

prompt-guard bundles 600+ offline security patterns for ai agent runtime defense, covering prompt injection, supply chain attacks, memory poisoning, and unicode steganography, but lacks actionable procedure detail.

ln-621-security-auditor scans code for hardcoded secrets, sql injection, xss, insecure dependencies, and missing input validation. returns structured findings with severity levels and compliance scoring.

ace-step wraps stepfun-ai's music foundation model via runcomfy cli for generation, inpainting, and outpainting. supports tag-driven workflows across four endpoints at 0.0002-0.0003 per second of audio, capped at 4 minutes per call.

defi-protocol-templates provides solidity code samples for staking and amm contracts but lacks procedural guidance, failure mode analysis, and integration instructions. suitable only as a code reference, not a skill.

base44-troubleshooter provides log fetching and filtering for production issue diagnosis within base44 projects, covering authentication, function-level drilling, and time-range correlation.

mutation-testing provides trigger patterns and scope bounds for configuring mewt/muton campaigns across multiple languages, but lacks substantive procedural steps, failure handling, or concrete workflow examples.

decay detects and flags stale evidence in decisions, enabling time-aware risk governance through expiry tracking and waiver workflows.

competitor-tracking monitors rival apps for ongoing changes in keywords, ratings, metadata, and features. it establishes a watchlist, defines review cadence, and produces change logs with alerts rather than static snapshots.

llmfit matches locally available hardware against hundreds of llm model profiles, recommending which models will execute efficiently given detected cpu, ram, and gpu resources. supports multiple runtime backends and quantization strategies.