AI Agent Skills

This skill covers OpenClaw-specific security topics across its full lifecycle: runtime sandboxing, permission and capability models, secrets handling, supply-chain protections for skills and the ClawHub registry, vulnerability disclosures, hardening practices, security-relevant features and patches, audit mechanisms, and incident reports involving OpenClaw or its agent runtime. OpenClaw has been implicated in multiple supply-chain incidents — the Clinejection attack via Cline CLI (February 2026, GHSA-9ppg-jx86-fqw7), a broader axios npm compromise distributing a cross-platform RAT via OpenClaw-impersonating packages (March 2026, attributed to Sapphire Sleet), and malicious skills on ClawHub — as well as documented architectural vulnerabilities including reasoning-text leaks to end-user channels, exec-approval resets on upgrade, susceptibility to prompt injection, memory poisoning, skill poisoning, and intent drift. Academic research (Tsinghua University / Ant Group, arXiv:2603.11619) has formally characterized these as multi-stage systemic threats and proposed a five-layer lifecycle-oriented defense framework, with the ClawAegis plugin released as a purpose-built implementation. Key hardening actions include: upgrading to at least v2026.3.8 (patches CVE-2026-33574 and 12+ other issues), v2026.4.25 (patches a privilege escalation in token rotation), and v2026.4.27 (adds Codex Computer Use with fail-closed pre-checks); configuring the two-layer exec-approval policy after every upgrade; explicitly suppressing reasoning output for production deployments; auditing transitive dependency trees for compromised packages; and treating all third-party ClawHub skills as untrusted executable code requiring review before installation. Enterprise alternatives such as PokeeClaw and Clawdi address native gaps in user isolation, RBAC, audit logging, and sandboxed execution that stock OpenClaw does not provide.

This skill covers actionable crypto and digital asset trading intelligence across the full spectrum: Bitcoin and altcoin price action, on-chain analytics (whale flows, exchange inflows/outflows), derivatives data (funding rates, open interest, liquidations), DeFi TVL dynamics, stablecoin supply and regulation, mining/hashrate trends, token unlocks, and cross-market macro correlations. It supports systematic and quantitative crypto strategy design — including factor selection, Sharpe ratio benchmarking, execution infrastructure requirements, and backtest-vs-live performance discipline — and includes evaluation of social-media trade calls using independent aggregator verification. Coverage extends to regulatory developments (CLARITY Act stablecoin legislation, Kraken/Bitnomial acquisition, HBAR ETF flows, Fed nominee crypto exposure, Pentagon Bitcoin disclosures), market structure events (RAVE manipulation and liquidation cascade, CME gap levels, DEX market share growth, Hyperliquid's geopolitical price discovery role, HIP-4 prediction market rollout), institutional adoption signals (UAE sovereign wealth Bitcoin ETF holdings, Morgan Stanley E*TRADE/MSBT launch, Coinbase UK USDC lending, RWA tokenization via Flow Capital/DigiFT, Decibel/Aptos perp DEX), on-chain regime signals (MVRV, Realized Price), and emerging infrastructure (Hyperliquid HIP-3/HIP-4, HYPE OTC accumulation and ETF filings, HypeStrat/Paradigm/Multicoin flows, Hyperbeat Liquid Banking, NEAR Intents cross-chain swaps, pre-IPO perpetuals). Always distinguish simulated pre-cost performance from live after-fee results, and cross-verify prices and sentiment readings from multiple independent sources before citing.

This skill covers actionable trading intelligence for traditional equity markets as of mid-to-late April 2026, including index movements (S&P 500, Nasdaq-100, Dow Jones, Russell 2000), sector rotations (defense, SaaS, semiconductors, financials), earnings reaction patterns (notably Netflix's repeated guidance-driven selloffs), mega-cap revenue benchmarks, Fed liquidity operations and balance sheet mechanics, geopolitical macro risk from the US-Iran conflict and Hormuz oil shock, institutional positioning dynamics (FOMO-driven flows per Barclays), options market structure (GEX, SpotGamma TRACE, 0DTE, Call Wall), and technical methodologies (TTM Squeeze, failed breakdowns, target slates, breakout confirmation filters). It also tracks specific catalysts including TSMC's Q1 2026 AI-demand beat, bank trading desk records, the SaaS sector dislocation, the Anthropic/OpenAI IPO pipeline, notable analyst calls (UNH Top Pick), and key insider/ownership filings (Uber/LCID). All guidance is observational intelligence only — not financial advice — and all price levels, ratings, and positioning data must be verified against live sources before acting.

This skill covers practical prompt engineering techniques and patterns for AI and LLM models, including named frameworks (RAIN, CLAR, FLOW, PIVO, SEED, RTF, RISE, RISEN, RHODES, etc.), chain-of-thought, few-shot prompting, role prompting, system prompt design, structured output, token optimization, and prompt injection mitigations. It includes model-specific guidance for Claude, GPT, Gemini, and open-source models, as well as tool-specific prompt shortcuts for Claude Code (slash commands, custom skills, plugins, and session management). The skill also covers writing system prompts for LLM-powered agents in automation platforms like n8n, including backend selection, tool enumeration, guardrails, memory configuration, and production reliability patterns. Broader context on the AI ecosystem — Anthropic's infrastructure, agentic email pipelines, Anthropic Academy, and the AI coding tools market — is included as supporting background. Always distinguish between model-level prompting (applicable in any interface) and tool-specific commands (applicable only in a particular IDE or CLI).