prompt-engineer_skill

This skill immediately invokes the prompt-engineering workflow via a Python script to optimize prompts without preliminary analysis, accelerating scope-driven
  • Python

538

GitHub Stars

3

Bundled Files

2 months ago

Catalog Refreshed

4 months ago

First Indexed

Readme & install

Copy the install command, review bundled files from the catalogue, and read any extended description pulled from the listing source.

Installation

Preview and clipboard use veilstrat where the catalogue uses aiagentskills.

npx veilstrat add skill solatis/claude-config --skill prompt-engineer

  • CLAUDE.md696 B
  • README.md8.1 KB
  • SKILL.md1.2 KB

Overview

This skill immediately runs a Python script to perform prompt optimization workflows when invoked. It is designed to be called directly (no pre-analysis) and follows a step-based process that triages scope then executes targeted optimization steps.

How this skill works

On invocation the skill runs the optimize script in the .claude/skills/scripts directory with a --step argument. Step 1 performs triage and returns a scope, then steps 2–6 apply the workflow for the chosen scope (single-prompt, ecosystem, greenfield, or problem). The script drives all inspection, edits, and outputs — follow its prompts and outputs rather than performing independent analysis first.

When to use it

  • User requests prompt optimization and expects immediate script execution
  • You need a triage-first workflow to determine scope before editing
  • Optimizing a single prompt, multiple interacting prompts, or designing a new prompt
  • Fixing a specific prompt issue identified by testing or user report
  • Automating a repeatable, scripted prompt-engineering process

Best practices

  • Invoke the skill with step 1 first to get an accurate scope determination
  • Pass the returned scope to subsequent steps 2–6 exactly as provided
  • Run the script from the .claude/skills/scripts working directory to ensure relative paths resolve
  • Provide clear problem descriptions or example inputs when using the problem scope
  • Do not perform external analysis before invoking the script — let the workflow run immediately

Example use cases

  • Run step 1 to classify a prompt into single-prompt vs ecosystem, then run step 2 with that scope
  • Use greenfield scope to produce a new prompt from product requirements provided to the script
  • Invoke problem scope when a prompt returns biased or incorrect outputs and follow the script’s repair steps
  • Optimize a set of related prompts (ecosystem) to reduce contradictions and improve interoperability
  • Automate repeated optimization runs as part of an integration that calls the script with different inputs

FAQ

Yes. Step 1 triages the request and determines the correct scope for steps 2–6.

What working directory should I use?

Invoke the script from .claude/skills/scripts so relative resources and modules load correctly.

Built by
VeilStrat
AI signals for GTM teams
© 2026 VeilStrat. All rights reserved.All systems operational