aiskillstore/marketplace / skills/89jobrien/prompt-optimization
prompt-optimization
VerifiedExpert prompt optimization for LLMs and AI systems. Use when building AI features, improving agent performance, crafting system prompts, or optimizing LLM interactions. Masters prompt patterns and techniques.
| File | Kind | Size |
|---|---|---|
| SKILL.md | skill md | 3.0 KB |
| skill-report.json | config | 15.0 KB |
skill-report.json
View on GitHub{
"schema_version": "2.0",
"meta": {
"generated_at": "2026-07-05T13:26:23.445Z",
"slug": "89jobrien-prompt-optimization",
"source_url": "https://github.com/89jobrien/steve/tree/main/steve/skills/prompt-optimization",
"source_ref": "main",
"model": "codex",
"analysis_version": "3.0.0",
"source_type": "community",
"content_hash": "1c8dfe31666de4a94f6046ac0ac600e801be0542ef980e406b89366b0c949bad",
"tree_hash": "b09bcab4504512ed520c335580489da27a633be8606ea88577953a3eae7561a5",
"provenance": {
"model_effective": "codex",
"fallback_chain": [
{
"agent": "codex",
"outcome": "success"
}
],
"analysis_version": "3.0.0"
}
},
"skill": {
"name": "prompt-optimization",
"description": "Expert prompt optimization for LLMs and AI systems. Use when building AI features, improving agent performance, crafting system prompts, or optimizing LLM interactions. Masters prompt patterns and techniques.",
"summary": "Optimize LLM prompts with clear structure, examples, constraints, and output formats.",
"icon": "π¦",
"version": "1.0.2",
"author": "Joseph OBrien",
"license": "MIT",
"category": "productivity",
"tags": [
"prompt-engineering",
"llm",
"ai-agents",
"few-shot-learning",
"system-prompts"
],
"supported_tools": [
"claude",
"codex",
"claude-code"
],
"risk_factors": [
"external_commands"
]
},
"security_audit": {
"risk_level": "safe",
"is_blocked": false,
"safe_to_publish": true,
"analysis_status": "ok",
"summary": "The static analyzer flagged Markdown code fences as Ruby or shell backtick execution. Review of SKILL.md shows these are prompt examples and fenced Markdown blocks, not executable commands. No prompt injection attempt or malicious intent was found.",
"remediation": [
{
"issue": "Markdown fences were misclassified as shell execution.",
"suggestion": "Label example fences with plain text or markdown languages to reduce false static alerts.",
"severity": "low"
}
],
"risk_factor_evidence": [
{
"factor": "external_commands",
"evidence": [
{
"file": "SKILL.md",
"line_start": 40,
"line_end": 42
},
{
"file": "SKILL.md",
"line_start": 42,
"line_end": 44
},
{
"file": "SKILL.md",
"line_start": 44,
"line_end": 46
},
{
"file": "SKILL.md",
"line_start": 46,
"line_end": 50
},
{
"file": "SKILL.md",
"line_start": 50,
"line_end": 52
},
{
"file": "SKILL.md",
"line_start": 52,
"line_end": 92
},
{
"file": "SKILL.md",
"line_start": 92,
"line_end": 96
},
{
"file": "SKILL.md",
"line_start": 96,
"line_end": 116
},
{
"file": "SKILL.md",
"line_start": 116,
"line_end": 123
}
]
}
],
"static_findings": [
{
"id": "external_commands:SKILL.md:40:ruby-shell-backtick-execution",
"category": "external_commands",
"severity": "medium",
"pattern": "Ruby/shell backtick execution",
"file": "SKILL.md",
"line_start": 40,
"line_end": 42,
"snippet": "```"
},
{
"id": "external_commands:SKILL.md:42:ruby-shell-backtick-execution",
"category": "external_commands",
"severity": "medium",
"pattern": "Ruby/shell backtick execution",
"file": "SKILL.md",
"line_start": 42,
"line_end": 44,
"snippet": "```"
},
{
"id": "external_commands:SKILL.md:44:ruby-shell-backtick-execution",
"category": "external_commands",
"severity": "medium",
"pattern": "Ruby/shell backtick execution",
"file": "SKILL.md",
"line_start": 44,
"line_end": 46,
"snippet": "```"
},
{
"id": "external_commands:SKILL.md:46:ruby-shell-backtick-execution",
"category": "external_commands",
"severity": "medium",
"pattern": "Ruby/shell backtick execution",
"file": "SKILL.md",
"line_start": 46,
"line_end": 50,
"snippet": "```"
},
{
"id": "external_commands:SKILL.md:50:ruby-shell-backtick-execution",
"category": "external_commands",
"severity": "medium",
"pattern": "Ruby/shell backtick execution",
"file": "SKILL.md",
"line_start": 50,
"line_end": 52,
"snippet": "```"
},
{
"id": "external_commands:SKILL.md:52:ruby-shell-backtick-execution",
"category": "external_commands",
"severity": "medium",
"pattern": "Ruby/shell backtick execution",
"file": "SKILL.md",
"line_start": 52,
"line_end": 92,
"snippet": "```"
},
{
"id": "external_commands:SKILL.md:92:ruby-shell-backtick-execution",
"category": "external_commands",
"severity": "medium",
"pattern": "Ruby/shell backtick execution",
"file": "SKILL.md",
"line_start": 92,
"line_end": 96,
"snippet": "```markdown"
},
{
"id": "external_commands:SKILL.md:96:ruby-shell-backtick-execution",
"category": "external_commands",
"severity": "medium",
"pattern": "Ruby/shell backtick execution",
"file": "SKILL.md",
"line_start": 96,
"line_end": 116,
"snippet": "```"
},
{
"id": "external_commands:SKILL.md:116:ruby-shell-backtick-execution",
"category": "external_commands",
"severity": "medium",
"pattern": "Ruby/shell backtick execution",
"file": "SKILL.md",
"line_start": 116,
"line_end": 123,
"snippet": "```"
}
],
"finding_verdicts": [
{
"id": "external_commands:SKILL.md:40:ruby-shell-backtick-execution",
"verdict": "false_positive",
"confidence": 0.96,
"reason": "Line 40 starts a Markdown code fence around a plain prompt example. It is documentation text, not Ruby backtick execution or a shell command."
},
{
"id": "external_commands:SKILL.md:42:ruby-shell-backtick-execution",
"verdict": "false_positive",
"confidence": 0.96,
"reason": "Line 42 closes a Markdown code fence for a prompt example. There is no executable context or command invocation."
},
{
"id": "external_commands:SKILL.md:44:ruby-shell-backtick-execution",
"verdict": "false_positive",
"confidence": 0.96,
"reason": "Line 44 opens another Markdown code fence containing a natural-language prompt. This is not code execution."
},
{
"id": "external_commands:SKILL.md:46:ruby-shell-backtick-execution",
"verdict": "false_positive",
"confidence": 0.96,
"reason": "Line 46 closes the fenced prompt example. The surrounding content is documentation with no interpreter or shell invocation."
},
{
"id": "external_commands:SKILL.md:50:ruby-shell-backtick-execution",
"verdict": "false_positive",
"confidence": 0.96,
"reason": "Line 50 is a Markdown fence before a prompt pattern example. It does not execute external commands."
},
{
"id": "external_commands:SKILL.md:52:ruby-shell-backtick-execution",
"verdict": "false_positive",
"confidence": 0.96,
"reason": "Line 52 closes a Markdown code fence after natural-language text. No shell syntax or command execution is present."
},
{
"id": "external_commands:SKILL.md:92:ruby-shell-backtick-execution",
"verdict": "false_positive",
"confidence": 0.97,
"reason": "Line 92 opens a Markdown-labeled example block. The block documents an optimized prompt and is not an executable script."
},
{
"id": "external_commands:SKILL.md:96:ruby-shell-backtick-execution",
"verdict": "false_positive",
"confidence": 0.97,
"reason": "Line 96 begins a nested prompt example inside documentation. The content is a sample system prompt, not shell execution."
},
{
"id": "external_commands:SKILL.md:116:ruby-shell-backtick-execution",
"verdict": "false_positive",
"confidence": 0.97,
"reason": "Line 116 closes the nested prompt example. It is a Markdown delimiter and has no command execution behavior."
}
],
"semantic_findings": [],
"critical_findings": [],
"high_findings": [],
"medium_findings": [],
"low_findings": [],
"dangerous_patterns": [],
"files_scanned": 1,
"total_lines": 142,
"audit_model": "codex",
"audited_at": "2026-07-05T13:26:23.445Z",
"agent_auto_install_policy": "allowed",
"manual_install_policy": "allowed"
},
"content": {
"user_title": "Optimize LLM Prompts and Agent Instructions",
"value_statement": "Weak prompts create inconsistent model outputs and slow AI feature development. This skill helps structure prompts, examples, constraints, and output formats for Claude, Codex, and Claude Code.",
"seo_keywords": [
"Claude",
"Codex",
"Claude Code",
"prompt optimization",
"prompt engineering",
"LLM prompts",
"system prompts",
"few-shot learning",
"AI agents",
"output formatting"
],
"actual_capabilities": [
"Defines prompt sections for role, task, constraints, output format, and examples.",
"Adds few-shot input and output examples for desired style and edge cases.",
"Converts vague requests into clear system prompts for agents and AI features.",
"Specifies response formats and evaluation criteria for repeatable outputs.",
"Suggests iteration practices for testing and refining prompts."
],
"limitations": [
"It does not test prompts against live model responses.",
"It does not provide model-specific benchmark scores or latency data.",
"It does not replace safety review for prompts that handle sensitive data.",
"Final quality depends on the provided task details and examples."
],
"use_cases": [
{
"title": "Design Agent Instructions",
"description": "Create structured system prompts for AI agents with roles, boundaries, and expected outputs.",
"target_user": "AI product builder"
},
{
"title": "Improve Review Workflows",
"description": "Turn broad review requests into clear prompts with criteria, severity labels, and action-focused outputs.",
"target_user": "Engineering lead"
},
{
"title": "Standardize Prompt Patterns",
"description": "Document reusable prompt structures for teams that need consistent model behavior across workflows.",
"target_user": "Technical writer"
}
],
"prompt_templates": [
{
"title": "Improve a Basic Prompt",
"scenario": "Beginner prompt cleanup",
"prompt": "Optimize this prompt for clarity, constraints, and output format: [paste prompt]. Explain the key changes in simple terms."
},
{
"title": "Create a System Prompt",
"scenario": "Structured agent instruction",
"prompt": "Create a system prompt for [agent or feature]. Include role, task, boundaries, inputs, output format, and two success criteria."
},
{
"title": "Add Few-Shot Examples",
"scenario": "Improve consistency with examples",
"prompt": "Rewrite this task prompt with three few-shot examples. Cover a typical case, an edge case, and a failure case."
},
{
"title": "Design an Agent Workflow",
"scenario": "Advanced multi-step prompt design",
"prompt": "Optimize this multi-step AI workflow for [goal]. Define agent roles, handoffs, verification checks, and prompt sections for each step."
}
],
"output_examples": [
{
"input": "Create a system prompt for a code review agent.",
"output": [
"Defines the reviewer role and scope.",
"Lists security, performance, maintainability, and best-practice checks.",
"Requires severity, evidence, explanation, and suggested fixes for each issue."
]
},
{
"input": "Improve a support triage prompt for inconsistent responses.",
"output": [
"Adds clear ticket categories and escalation rules.",
"Includes examples for urgent, normal, and incomplete requests.",
"Specifies a concise response format for routing decisions."
]
},
{
"input": "Add few-shot examples to a data labeling prompt.",
"output": "The optimized prompt includes representative labels, edge cases, rejection guidance, and a required explanation for uncertain decisions."
}
],
"best_practices": [
"Start with the user goal, available inputs, and required output before adding examples.",
"Include edge cases that show what the model should avoid or reject.",
"Test revised prompts with representative tasks before using them in production."
],
"anti_patterns": [
"Asking for better results without defining audience, constraints, or success criteria.",
"Mixing task instructions, examples, and formatting rules in one unstructured paragraph.",
"Treating a single good response as proof that a prompt is reliable."
],
"faq": [
{
"question": "What does this skill optimize?",
"answer": "It optimizes prompts for LLM tasks, AI agents, system instructions, examples, constraints, and output formats."
},
{
"question": "Can it create prompts for Claude and Codex?",
"answer": "Yes. It provides prompt structure that can be adapted for Claude, Codex, Claude Code, and similar LLM tools."
},
{
"question": "Does it run benchmarks?",
"answer": "No. It provides prompt design guidance, but it does not execute tests or collect benchmark results."
},
{
"question": "Does it support few-shot prompting?",
"answer": "Yes. It recommends input and output examples, including typical cases and edge cases."
},
{
"question": "Can it help with system prompts?",
"answer": "Yes. It can structure system prompts with roles, task boundaries, constraints, and required response formats."
},
{
"question": "Is this a safety review tool?",
"answer": "No. It can improve prompt clarity, but sensitive or high-risk workflows still need separate safety review."
}
]
},
"file_structure": [
{
"name": "SKILL.md",
"type": "file",
"path": "SKILL.md",
"lines": 142
}
]
}
Related skills
| Skill | Stars | Updated | Works with | Status | Install |
|---|---|---|---|---|---|
| reverse-engineer-anythingmorluto/reaReverse engineer native, managed, Electron/JavaScript, packaged, firmware, and browser targets with REA, and build agent environments (AgentEnv bundles) of web applications from a live instance or a capture, and of Linux desktop apps with a task from a request. Use shipped-artifact or requested runtime evidence to explain features, compare versions, decompile code, or guide a reconstruction. Skip REA for ordinary source-repository architecture analysis. | Stars: 472k | Updated | Claude CodeCodex | Verified | Install |
| diagnosing-superpowersobra/superpowersUse when a superpowers session went wrong and your human partner wants to know why β repeated work, ignored plans, stumbles, poor results, a skill that didn't fire, "it took too long", "why is it so expensive", "what is it doing" β or wants to build a bug report for the superpowers maintainers, for the current session or a past one identified by id or path, on any harness. | Stars: 322k | Updated | Claude CodeCodex | Verified | Install |
| dispatching-parallel-agentsobra/superpowersUse when facing 2+ independent tasks that can be worked on without shared state or sequential dependencies | Stars: 322k | Updated | Claude CodeCodex | Verified | Install |
| executing-plansobra/superpowersUse when executing an implementation plan in the current session as the implementer yourself β your human partner chose inline execution, or no subagent tool is available | Stars: 322k | Updated | Claude CodeCodex | Verified | Install |