@connor_dev1
Founder of Pergamum Building lots of random things
A senior-engineer-style code review. Cites line numbers, suggests concrete fixes, skips nitpicks unless they matter.
Turns a git diff into a structured PR description. Plain, honest, no fluff.
Explains a SQL query in plain English, identifies real performance issues, and rewrites it for your database.
Generates a production-ready API endpoint with validation, auth, error handling, and types — ready to paste in.
Turns a vague bug report into a triaged ticket with repro steps, severity, and a starting guess at the cause.
Turns raw interview notes into prioritized pain points, surprises, and one concrete next action.
Strips a messy meeting note into clean decisions, owned action items, and open questions.
Five different opening lines for a marketing email — five different angles, no clichés, each under 18 words.
Three personalized cold-email openers based on a real LinkedIn or website excerpt. No flattery, no fabrications.
Turns a one-line product idea into a real spec — user, problem, MVP scope, and what to skip for v1.
Turns a git log into clean release notes — highlights, new, improved, fixed, breaking. No 'we are thrilled.'
Turns a long-form blog post into a tight X thread that earns the click and invites the read.
Rewrites text in your brand voice without padding, without clichés, and without changing the meaning.
Turns a goal into proper OKRs — one ambitious objective, 3-4 measurable key results, with leading indicators.
Writes a full Architecture Decision Record — context, decision, alternatives, consequences, open questions.
Turns raw observations into a structured perf review — specific strengths, fixable growth areas, real calibration.
Writes a real job description — concrete responsibilities, soft requirements, transparent comp. No 'Ninja' titles.
Writes proper docs for a function or API endpoint — summary, params, returns, examples, real gotchas.
Turns a topic into a real content brief — title options, hook, outline, sources, and the distinctive angle.
Plans a refactor against a specific constraint — small reviewable steps, real dependencies, what to skip.
> **How to use:** Fill in all `{{variables}}` below, then paste this entire prompt into Claude, ChatGPT GPT, or your AI tool of choice. Paste your codebase, repo contents, or specific files at the bottom. For best results, run it per layer of the stack (frontend, backend, API, infrastructure) and compile the outputs into a single findings report.
How to use:** Fill in all `{ {variables} }` below, then paste this entire prompt into Claude, ChatGPT GPT, or your AI tool of choice. Paste your codebase, repo contents, or specific files at the bottom. For best results, run it per layer of the stack (frontend, backend, API, infrastructure) and compile the outputs into a single findings report.
> **How to use:** Fill in the 8 variables, then run this prompt in Claude, ChatGPT, or your AI tool of choice. Paste your codebase, repo contents, or specific files at the bottom. For best results, run it per layer of the stack (frontend, backend, API, infrastructure) and compile the outputs into a single findings report.
A senior-engineer sign-off pass that ranks the real problems worst first.
Flag risky, outdated, unused, or vulnerable dependencies and what to swap.
Find real vulnerabilities ranked by severity and get the exact fix for each.
Scan code for one specific thing and list every instance with a fix.
Turn a feature into small build-and-verify steps you can run with an AI coding tool.
Generate tests that would actually catch bugs, not trivial filler.
Find the real bottlenecks worst first and fix the parts that matter.