发现 Skills

Create a new Hydra agent with full configuration. Use when user wants to add an agent, new agent, create agent, or set up another agent. Walks through persona, mounts, Telegram, model, and secrets.

Workflow for preventing secret leaks and sensitive logging (PII/credentials) and adding redaction defaults.

Audit Kubernetes RBAC, enforce policies, and manage secrets. Use for security reviews, permission audits, policy enforcement with Kyverno/Gatekeeper, and secret management.

Break down complex tasks into small, manageable, atomic units

Static application security testing (SAST) using Semgrep for vulnerability detection, security code review, and secure coding guidance with OWASP and CWE framework mapping. Use when: (1) Scanning code for security vulnerabilities across multiple languages, (2) Performing security code reviews with pattern-based detection, (3) Integrating SAST checks into CI/CD pipelines, (4) Providing remediation guidance with OWASP Top 10 and CWE mappings, (5) Creating custom security rules for organization-specific patterns, (6) Analyzing dependencies for known vulnerabilities.

Converts JSON data snippets into Python Pydantic data models.

Generates optimized read queries using Dapper. Includes multi-mapping for joins, pagination, dynamic filtering, CTEs, and best practices for high-performance data access.

Generative Engine Optimization for AI search engines (ChatGPT, Claude, Perplexity).

Discover new feature opportunities from multiple sources — creative brainstorming, user feedback, ecosystem evolution, technical possibilities, and cross-project research — and add them to .feature-radar/opportunities/. Use this when the user wants to generate NEW ideas rather than evaluate existing ones. Even casual mentions like "I wonder what else we could do" or "let's think about new features" should trigger this skill. Use when: - Brainstorming creative feature ideas or exploring new directions - User says "what else could we build?" or "give me some ideas" - Reviewing user feedback, GitHub issues, or community requests for unmet needs - Exploring adjacent tools and ecosystems for inspiration - Periodic opportunity refresh to keep the backlog current Trigger phrases: "scan opportunities", "find new features", "brainstorm ideas", "what could we build", "refresh opportunities", "scan ecosystem", "give me feature ideas", "what are we missing", "explore new directions"

Production-first enterprise skill for The Composable Architecture (TCA) with SwiftUI (iOS 16+, TCA 1.7+). Use this skill when building new TCA features with the @Reducer macro, decomposing fat/god reducers, implementing StackState/StackAction navigation or tree-based @Presents navigation, writing TestStore tests, migrating legacy TCA code to modern @ObservableState patterns, debugging TCA performance issues, managing side effects and dependencies with @DependencyClient, or reviewing TCA code for anti-patterns. Apply it whenever working with TCA reducers, stores, effects, or dependencies — AI tools often generate outdated pre-1.7 TCA patterns, so this skill is essential to produce correct, production-ready code.

Cache Playwright browser binaries in CI/CD pipelines (GitHub Actions, Azure DevOps) to eliminate 1-2 minute download overhead on every build. Uses version-based cache keys derived from Central Package Management (CPM) for automatic invalidation. Use when setting up or optimizing CI/CD for .NET projects with Playwright E2E tests.

Guide for writing effective technical documentation, specs, and markdown files. Use when creating or editing specs, documentation, READMEs, AGENTS.md files, or any markdown content in the codebase.

Use when writing code, reviewing PRs, refactoring, or making architectural decisions. Enforces readability, testability, and maintainability standards across all languages.

Cloudflare Workers performance optimization with CPU, memory, caching, bundle size. Use for slow workers, high latency, cold starts, or encountering CPU limits, memory issues, timeout errors.

Automated security scanning for dependencies, code, containers with Trivy, Snyk, npm audit. Use for CI/CD security gates, pre-deployment audits, compliance requirements, or encountering CVE detection, outdated packages, license compliance, SBOM generation errors.

Manages automated **dynamic/trailing** stop losses (DSL only) for leveraged perpetual positions on Hyperliquid. Monitors price via cron, ratchets profit floors through configurable tiers, and auto-closes positions on breach via mcporter — no agent intervention for the critical path. Supports LONG and SHORT, strategy-scoped state isolation, and automatic cleanup on position or strategy close. ROE-based (return on margin) tier triggers that automatically account for leverage. Use only when the user wants a **trailing/dynamic** stop loss (DSL). Do not use for normal/static stop loss. If the user says "stop loss" without specifying DSL vs normal, ask which they mean before proceeding.

SENTINEL v1.0 — Quality Trader Convergence Scanner. Inverted pipeline: finds assets where smart-money (SM) is accelerating FIRST, then checks who's profiting from them using momentum-event TCS/TRP quality tags. Catches the trade BEFORE top traders hit the leaderboard, not after. Three confirmation layers: rising SM contribution velocity + quality traders confirmed via momentum events + optional top-trader cross-check. 2–4 trades per day, highest conviction in the zoo.

Developing, testing, and deploying Streamlit data applications on Snowflake. Use this skill when you're building interactive data apps, setting up local development environments, testing with pytest or Playwright, or deploying apps to Snowflake using Streamlit in Snowflake.

Launch the Gemini CLI as a headless sub-agent in a tmux pane, monitor its stream-json JSONL output, and poll for completion. Use this when delegating a well-defined task to a Gemini agent subprocess with live observability. Triggers include "run gemini agent", "delegate to gemini", "gemini sub-agent", or any request to run Gemini headlessly and monitor it.

Multi-agent collaborative planning system that automatically decomposes complex tasks into parallelizable sub-tasks, manages agent orchestration through DAG scheduling, and provides comprehensive observability. Extends planning-with-files with intelligent agent activation, dependency management, and context isolation to solve context overload problems.

Generate images and videos using fal.ai AI models

[Extended thinking: This workflow implements a sophisticated debugging and resolution pipeline that leverages AI-assisted debugging tools and observability platforms to systematically diagnose and resolve incidents.]

Automate Instagram tasks via Rube MCP (Composio): create posts, carousels, manage media, get insights, and publishing limits. Always search tools first for current schemas.

Provide proven marketing strategies and growth ideas for SaaS and software products, prioritized using a marketing feasibility scoring system.