AGENTKIT/WORKS

pack · tripwire tier

Prompt Ops: 150 Production Prompts

150 production prompts, annotated — the failure modes documented for every one.

PRICE$29$19
Buy now — $19

Exactly 150 prompts — counted in the manifest, not rounded — across debugging, refactoring, code review, testing, spec writing, research, incident response, documentation, and data analysis. Every prompt annotated with when to use it, what to expect, and the failure modes to watch for. No 'act as a world-class expert' filler.

$ tree prompt-ops-150/✓ actual contents
prompt-ops/
├── README.md · LICENSE
├── CHAINING.md          ← 6 pipeline patterns
├── prompts.json         ← whole library, machine-readable
└── prompts/
    ├── 01-debugging.md          (17)
    ├── 02-code-review.md        (17)
    ├── 03-refactoring.md        (17)
    ├── 04-testing.md            (17)
    └── … 5 more categories      (82)

Why it's different

  • ▸ Failure-mode notes on every prompt: when it will mislead the model, and what to do instead
  • ▸ 6 chaining patterns for multi-step work — the part you can't get by asking ChatGPT for 'a good prompt'
  • ▸ prompts.json ships alongside the Markdown — pipe the whole library into your own tooling

Questions, answered

Why buy prompts when I can write my own?+

You're not paying for the prompts — you're paying for the annotations. Each one documents the failure mode we hit using it, and the 6 chaining patterns show how prompts hand off to each other in multi-step work. That's weeks of burned sessions, written down.

Which models do these target?+

Written for Claude-class frontier models but deliberately model-agnostic — the annotations explain the intent so you can adapt them anywhere.

How do updates work?+

Buy once, keep it. New versions ship free for 12 months — you'll get an email when one drops, and a reply to your receipt gets a fresh link any time.

Kits that pair with this one

Each workflow kit teaches a different job. They share the same skill standard, so they stack.