LLM Mart Trusted

A Claude skill that reviews a diff by trying to refute each finding before reporting it.

A prompt that forces Gemini to cite sources and flag uncertainty instead of guessing.

A workflow that makes DeepSeek state a hypothesis and a test before proposing any fix.

A Cursor recipe that writes a characterization test before each refactor step.

A skill that keeps a consistent persona and memory across a long MiniMax conversation.

Turns a staged diff into a clean Conventional Commits message — correct type, tight subject, useful body.

Uses ChatGPT search plus an adversarial verification pass to turn a fuzzy question into a sourced brief you can hand to someone else.

Forces Codex CLI to show the diff, review its own changes, run focused validation, and list uncertainty before you keep the patch.

A review skill that makes Copilot focus on concrete behavior changes, missing tests, and repo instructions instead of generic style chatter.

A starter layout for .github/copilot-instructions.md, path-specific instruction files, and AGENTS.md so Copilot stops guessing how your repo works.

Use /setupTests, /tests, and /fixTestFailure in a red-green loop instead of asking Copilot to “add some tests” and hoping for the best.

Forces a failing test that reproduces the bug before any fix — so you know it's actually fixed.

A reusable test harness for checking whether a custom GPT actually follows its instructions, uses tools correctly, and fails safely.

A structured security pass over a diff — checks the OWASP-relevant classes, ranks by exploitability, and shows a repro.

Stops the model from coding too early — lock a short spec and a test list before a single line is written.

The whole public catalogue is an MCP server and a REST API, so your agent can search skills, tools and slash-commands as native tools. Setup is one config block. Plus a private vault that carries your own prompts between machines.

You wouldn't ship a function you ran once. Here's the smallest evaluation setup that catches real regressions — five cases, three graders, and a rule for when to add a sixth.

AGENTS.md, CLAUDE.md, copilot-instructions.md and .cursor/rules all solve the same problem. Most of them are written badly. Here's what changes agent behaviour and what's decoration.

Sanitizing output protects your page. It does nothing for an agent that reads a poisoned README and then runs a command. A practical model of the threat, and what actually helps.

The words are used interchangeably and they shouldn't be. Six categories, what actually distinguishes them, and how to tell which one you're holding.

Most prompts are lucky one-offs. The ones that keep working share a structure: an explicit role, a checklist of constraints, a worked example, and a self-check step. Here's how to build one.

Set the model effort/reasoning level.

Rewind the conversation and/or code to a previous checkpoint.

Switch to planning mode and optionally send an inline planning prompt.

Condense the text into two or three sentences.

Discourage shallow, generic output.

Show the debug log path and copy it to the clipboard.

Ask the AI to analyze the current agent debug logs for troubleshooting.

Fan out web searches, cross-check sources, and synthesize a cited report.

Re-scan skill and command directories for changes made on disk.

Generate an AGENTS.md scaffold in the current directory for persistent repo instructions.

Return ready-to-use code with comments in the given language.

Open Settings to adjust theme, model, output style, and other preferences.

Discover features through quick interactive lessons with animated demos.

Toggle voice dictation, or enable it in a specific mode.

Analyze an attached PDF, document, or spreadsheet.

Enable or disable experimental Codex features from the CLI.

Rewrite the text in a specific style or platform format.

Configure sandbox mode and network access.

Ask a quick side question without adding to the conversation history.

Edit CLAUDE.md memory files and manage auto-memory.

OpenAI's general-purpose assistant for writing, research, coding, data analysis, and agentic tasks.

GitHub's AI pair programmer for inline suggestions, chat, agent mode, and code review across your editor and GitHub.

Anthropic's agentic CLI that edits code, runs tools, and ships changes.

The AI-first code editor — chat, edit, and refactor across your whole repo.