Prompt engineer
Skill yuri-semenenko/ai-engineering-workspace/codex/skills/prompt-engineer
One engineering workflow across Claude Code, Codex, Copilot, and Gemini CLI. A portable persona canon, process skills, and safety guardrails, kept in sync by design.
npx -y skills add yuri-semenenko/ai-engineering-workspace --skill prompt-engineerAssembled from the repository path, not quoted from the project. Check it against their README if it does not work.
One thing to look at
- 0 stars0 stars. Stars are a popularity signal and not a quality one, but at this level it is likely that nobody has read this closely except its author, and you would be relying on your own review.
What its author says it does
Copied from the file, not written here
Use when creating, rewriting, or reviewing prompts for AI agents, copilots, reusable workflows, RFC/review prompts, or ambiguous tasks that need stronger context, constraints, success criteria, and output structure.
SKILL.md
1.6 KB, 285 tokens by cl100k_base, as published. Nobody here has run it
Prompt Engineer
Turn rough intent into a structured, decision-oriented prompt for an AI agent.
Principles
- Structure prompts as goal, context, task, constraints, output format, and success criteria.
- Make assumptions and boundaries explicit.
- Use concrete examples only when they clarify behavior.
- Keep prompts token-efficient.
- Avoid generic role filler such as "You are a helpful assistant."
- Make the prompt strict enough to prevent drift and open enough to allow judgment.
Output Structure
Generate prompts with these sections:
- Role — what kind of agent should answer.
- Context — what the agent must know.
- Task — the concrete job.
- Constraints — required behavior and forbidden behavior.
- Output Format — exact response shape.
- Success Criteria — how the result will be judged.
- Edge Cases / Anti-patterns — common failure modes to avoid.
Mirror the user's language in the surrounding response. Generate the prompt itself in English unless the user asks otherwise.
Quality Bar
- Drive clarity, decisions, risks, dependencies, and ownership.
- Include assumptions and explicit non-goals.
- Ask 1-2 clarifying questions if missing context would make the prompt brittle.
- Defer to
rfc,adr, orpr-classifywhen the user wants that artifact directly.