wtf is prompt-ops-maker?

verisworks-ai/prompt-ops-maker — explained in plain English

Analysis updated 2026-05-18

★ 27PythonAudience · developerComplexity · 2/5LicenseSetup · easy

Mindmap

mindmap
  root((repo))
    What it does
      Generates agent prompts
      Adds verification gates
      Adds deny lists
    Tech stack
      Python
      CLI tool
      YAML config
    Use cases
      Build project specific prompts
      Analyze existing prompts
      Standardize agent operating rules
    Audience
      AI agent builders
      DevOps teams
    Analysis
      Local rule checks only
      No AI call
      No secret echo

mindmap root((repo)) What it does Generates agent prompts Adds verification gates Adds deny lists Tech stack Python CLI tool YAML config Use cases Build project specific prompts Analyze existing prompts Standardize agent operating rules Audience AI agent builders DevOps teams Analysis Local rule checks only No AI call No secret echo

Why would anyone build with this?

REASON 1

Generate a reusable operating prompt for an AI agent before it starts a long task

REASON 2

Score an existing prompt for missing safety rules like deny lists or verification gates

REASON 3

Build a one off ad-hoc prompt for a task without a saved project config

REASON 4

Produce structured JSON or YAML prompt analysis output for automated pipelines

What's in the stack?

PythonCLIYAMLMCP

How it stacks up

	verisworks-ai/prompt-ops-maker	avbiswas/sam2-mlx	chanthruu/wind_energy_monitor
Stars	27	27	27
Language	Python	Python	Python
Setup difficulty	easy	moderate	moderate
Complexity	2/5	4/5	3/5
Audience	developer	researcher	data

Figures from each repo's GitHub metadata at analysis time.

How do you spin it up?

Difficulty · easy Time to first run · 5min

Installs via pip from a GitHub clone, no external services or API keys needed for prompt generation.

MIT license: use, copy, modify, and distribute freely, provided as is with no warranty.

Wtf does this do

prompt-ops-maker is a command line tool that writes operating instructions for AI agents like Claude, Codex, Hermes, or Gemini. The idea behind it is that AI agents work better and safer when their prompt spells out what they are allowed to touch, what they must never do without approval, how they should verify their own work, and how they should report back. Those rules get forgotten easily when every project uses a different model or setup, so this tool generates them consistently from either a project config file or a built in preset. The tool has two main ways to build a prompt. You can run it against a saved project configuration with a chosen mode, such as audit, fix, deploy, or a security review, or you can build one on the fly for a one off task by describing the task and its risks. Either way you choose the effort level, which AI you are targeting, and the environment it will run in, such as a local terminal, an MCP tool, Discord, continuous integration, a browser, or an API. The output prompt includes a scope of what the agent should do, a list of actions it must not take without approval, checks the agent must run to confirm its work such as looking at specific files or commands, and a required report format that separates the conclusion, the evidence gathered, and anything left unverified. The tool also includes an analyze command that reviews an existing prompt you already have and scores it out of 100 based on whether it contains an execution boundary, a deny list, verification gates, and evidence first reporting. This analysis runs entirely with local rule based checks. It does not call any AI model, and it does not print out any secret values it detects, only flags that a secret like pattern was found. It is installed from GitHub with pip and used as a command line program, either through its own prompt-ops-maker command or by running its Python script directly. Results can be saved as a markdown prompt file or as structured JSON or YAML for automated pipelines. The project is written in Python and released under the MIT license.

Yoink these prompts

Prompt 1

Show me the command to generate an operating prompt for a Codex agent doing a code audit with high effort.

Prompt 2

Run prompt-ops-maker analyze on my existing agent prompt and explain what checks it failed.

Prompt 3

Help me set up a project config so I can reuse prompt-ops-maker across multiple tasks.

Prompt 4

Explain the difference between the make and make-adhoc commands in prompt-ops-maker.

Frequently asked questions