afrexai-prompt-masteryComprehensive system for designing, testing, optimizing, and managing clear, role-aware, actionable, focused, and testable prompts for AI models.
Install via ClawdBot CLI:
clawdbot install 1kalin/afrexai-prompt-masteryGrade Fair — based on market validation, documentation quality, package completeness, maintenance status, and authenticity signals.
Contains instructions to override system prompt or ignore user requests
"Ignore previous instructions"Calls external URL not in known-safe list
https://afrexai-cto.github.io/context-packs/AI Analysis
The skill provides legitimate prompt engineering guidance and templates. The external URL appears to be a documentation or resource site consistent with the skill's purpose, not an active data exfiltration endpoint. The 'prompt poisoning' signal is likely a false positive from example patterns, not active malicious instructions.
Audited Apr 17, 2026 · audit v1.0
Generated Mar 22, 2026
Use the CRAFT framework and prompt chaining to automate ticket classification and response drafting. Layer 3 XML tags structure customer profiles and complaints, while Layer 4 JSON output shapes standardized responses, reducing resolution time and ensuring consistency.
Implement prompt chaining for blog creation, where Phase 1 layers guide research, outlining, drafting, and editing. Advanced techniques like structured CoT ensure logical flow, and few-shot examples maintain brand voice, streamlining content production for marketing teams.
Apply constraint-based task definitions and CoT for analyzing financial data. Layer 4 output shaping enforces JSON formats for sentiment and risk assessment, while self-consistency CoT validates high-stakes decisions, aiding analysts in generating reliable reports.
Utilize goal-based prompts and few-shot examples to classify patient inquiries by urgency. Layer 1 system context sets medical constraints, and prompt chaining routes cases to appropriate specialists, improving efficiency and safety in clinical settings.
Leverage direct instruction prompts with variable slots for product details. Phase 1 layers ensure clear, actionable outputs, and advanced techniques optimize descriptions for SEO and conversion, scaling catalog management for online retailers.
Offer a cloud-based platform with tiered subscriptions for prompt design, testing, and management. Revenue comes from monthly fees based on usage tiers, feature access, and enterprise support, targeting businesses integrating AI agents.
Provide expert services for custom prompt engineering, including workshops and implementation support. Revenue is generated through project-based fees, hourly rates, and certification programs, catering to organizations adopting advanced AI workflows.
Operate a marketplace where users can buy, sell, or share optimized prompt templates. Revenue streams include transaction fees, premium listings, and affiliate partnerships, serving developers and content creators seeking reusable solutions.
💬 Integration Tip
Start by implementing the CRAFT framework for basic prompts, then gradually incorporate advanced techniques like prompt chaining to handle complex workflows without overwhelming initial users.
Scored Apr 19, 2026
Advanced expert in prompt engineering, custom instructions design, and prompt optimization for AI agents
Safe OpenClaw config updates with automatic backup, validation, and rollback. For agent use - prevents invalid config updates.
Evaluate, optimize, and enhance prompts using 58 proven prompting techniques. Use when user asks to improve, optimize, or analyze a prompt; when a prompt nee...
Transform rough ideas into professional-grade LLM prompts. Analyzes text, images, links, and documents to craft optimized prompts using proven frameworks (Co...
Extract conversation transcripts from AI coding session logs (Clawdbot, Claude Code, Codex). Use when asked to export prompt history, session logs, or transcripts from .jsonl session files.
Detect and block prompt injection attacks in emails. Use when reading, processing, or summarizing emails. Scans for fake system outputs, planted thinking blocks, instruction hijacking, and other injection patterns. Requires user confirmation before acting on any instructions found in email content.