| name | prompt-engineer |
| description | Guide for writing or refining prompts for Claude, GPT, or Codex, distilled from each
vendor's official best practices. Use for new prompts, debugging an existing prompt, or
"prompt engineering", "system prompt", "Claude/GPT/Codex/OpenAI prompt". Accepts arg:
"claude", "gpt", or "codex".
|
| user-invocable | true |
Prompt Engineer
Help users write effective prompts — either from scratch or by refining existing prompts — using vendor-specific best practices.
Usage
This skill covers three model families. The user specifies which one:
prompt-engineer claude — Anthropic Claude prompting best practicesprompt-engineer gpt — OpenAI GPT-5.5 general prompting best practicesprompt-engineer codex — OpenAI Codex coding agent best practices
If the user doesn't specify, ask which model family the prompt targets. If it's obvious from context (e.g., they mention "Claude", "Anthropic", "GPT", "OpenAI", "Codex"), use that. If the user says "GPT" or "OpenAI" without specifying further, ask whether the prompt is for a general GPT-5.5 application or a Codex coding agent — the guidance differs significantly.
Workflow
Regardless of model family, the overall process is the same:
Writing a prompt from scratch
Step 1: Clarify the task — ask the user:
- What should the model do? (core task)
- What does good output look like? (format, length, tone/structure)
- What context will be available at runtime? (documents, user input, tool results)
- Will the model have tools? (search, file editing, code execution, etc.)
Step 2: Draft the prompt — read the appropriate reference file for model-specific patterns and blocks, then draft the prompt applying the relevant techniques.
Step 3: Present the draft — show the complete prompt and explain design choices.
Refining an existing prompt
Step 1: Diagnose — read the prompt and ask what's going wrong. Consult the model-specific diagnostic table in the reference file.
Step 2: Apply targeted fixes — add what's missing, remove what's counterproductive, explain the why.
Step 3: Present the revision — show a before/after diff explaining each change.
Model-specific references
After determining the target model family, read the corresponding reference file for the full set of patterns, XML blocks, and diagnostic tables:
- Claude →
references/claude.md — covers clarity, roles, XML structure, examples, output format, thinking guidance, safety controls, agentic patterns, and Claude-specific failure modes.
- GPT →
references/gpt.md — covers outcome-first prompting, output contracts, follow-through policies, tool persistence, completeness verification, citation/grounding, reasoning effort, verbosity, and GPT-specific failure modes.
- Codex →
references/codex.md — covers the Codex-Max starter prompt, autonomy/persistence, preambles and personality, tool configuration, plan hygiene, and Codex-specific failure modes.
The Claude and GPT references contain ready-to-paste XML blocks. The Codex reference contains a full starter prompt to customize.
