From project-toolkit
Optimizes system prompts for Claude Code agents using research-backed prompt engineering patterns. Use for requests to improve, refine, or review agent workflows, tool instructions, or behaviors.
How this skill is triggered — by the user, by Claude, or both
Slash command
/project-toolkit:prompt-engineerclaude-sonnet-4-6The summary Claude sees in its skill listing — used to decide when to auto-load this skill
Optimizes system prompts by applying research-backed prompt engineering patterns. Human-in-the-loop phases: understand, plan, propose changes, receive approval, then integrate.
Optimizes system prompts by applying research-backed prompt engineering patterns. Human-in-the-loop phases: understand, plan, propose changes, receive approval, then integrate.
A well-optimized prompt achieves:
Optimization is complete when:
| Trigger Phrase | Operation |
|---|---|
optimize this prompt | Full Phase 0-4 optimization workflow |
improve this system prompt | Analyze and propose changes with visual cards |
review my agent prompt | Pattern-based review against reference |
refine this prompt for better results | Targeted improvement with BEFORE/AFTER |
make this prompt more effective | Technique selection and application |
Use when the user provides a prompt and wants it improved, refined, or reviewed for best practices.
Do NOT use for:
Before ANY analysis, read the appropriate pattern reference(s):
Read references/prompt-engineering-single-turn.md
Contains: Technique Selection Guide table, Quick Reference principles, domain-organized techniques with citations, Anti-Patterns section.
Read references/prompt-engineering-multi-turn.md
Read ONLY when the prompt involves:
Skip for:
Read references/workflow.md
Contains: Detailed Phase 0-4 workflows, visual card template, completion checkpoint.
┌─────────────────────────────────────────────────────────────────┐
│ 1. READ THE REFERENCE(S) │
│ - Always: references/prompt-engineering-single-turn.md │
│ - If multi-turn/multi-agent: also read multi-turn reference │
├─────────────────────────────────────────────────────────────────┤
│ 2. UNDERSTAND THE PROMPT (Phase 1) │
│ - Operating context (single-shot? tool-use? constraints?) │
│ - Current state (working? unclear? missing?) │
│ - Document specific problems with quoted prompt text │
├─────────────────────────────────────────────────────────────────┤
│ 3. PLAN WITH VISUAL CARDS (Phase 2) │
│ - Present each change as a visual card with: │
│ SCOPE → PROBLEM → TECHNIQUE → BEFORE/AFTER │
│ - Quote trigger conditions from reference │
│ - ⚠️ WAIT FOR USER APPROVAL before proceeding │
├─────────────────────────────────────────────────────────────────┤
│ 4. EXECUTE APPROVED CHANGES (Phase 3) │
│ - Apply the BEFORE → AFTER transformations │
├─────────────────────────────────────────────────────────────────┤
│ 5. INTEGRATE AND VERIFY QUALITY (Phase 4) │
│ - Check cross-section coherence │
│ - Final anti-pattern check │
│ - Present complete optimized prompt │
└─────────────────────────────────────────────────────────────────┘
Simple prompts (use lightweight process):
Complex prompts (use full process):
Before presenting the final prompt, verify:
| Avoid | Why | Instead |
|---|---|---|
| Applying techniques without reading reference first | Missing trigger conditions and constraints | Always read reference documents before analysis |
| Rewriting entire prompt | Destroys what already works | Preserve working sections, improve problems only |
| Skipping user approval before changes | May misidentify improvement priorities | Present visual cards in Phase 2, wait for approval |
| Stacking conflicting techniques | Produces contradictory instructions | Check stacking compatibility per reference |
| Using more than 3 emphasis markers | Dilutes signal when everything is emphasized | Reserve emphasis for highest-priority instructions |
After optimization:
npx claudepluginhub rjmurillo/ai-agents --plugin project-toolkitOptimizes system prompts for Claude Code agents using research-backed prompt engineering patterns. Use when users request prompt improvement, optimization, or refinement for agent workflows, tool instructions, or system behaviors.
Creates, optimizes, and iteratively refines agent prompts, system prompts, and reusable templates. Handles prompt tuning, reliability improvements, and porting between OpenAI, Claude, or Gemini.
Creates or improves production-grade system prompts for autonomous coding agents using evidence-gated workflows, explicit tool contracts, and completion criteria.