Posting guidelines
These guidelines exist so that a person arriving from a search engine finds a clear question, a clear answer and enough detail to judge whether it fits their own code. They are short on purpose. When something is not covered here, ask whether the post helps the next reader.
Stay on topic
The forum is for questions about code you did not write yourself: code produced by an assistant or agent, code copied from somewhere else, and the tools, rules and records around it. That includes prompts that produce bad code, questions about who or what wrote a piece of code, project and workplace policies on AI-written code, MCP servers, agent skills and coding agents, and the old scripts from the original library. General programming questions with no link to any of that are better asked elsewhere.
Asking a question
Put the real question in the title. "How do I stop my assistant adding a library I did not ask for?" is better than "Help with AI code". In the body, give the details we would otherwise have to ask for.
- The tool. Name the assistant, agent, editor plugin or MCP server, and its version or model where you know it.
- The prompt. Quote the prompt you used, word for word, in a code block. A paraphrase hides the line that caused the problem.
- The output. Quote what came back, or the part of it that is wrong. Trim it to what matters, but do not rewrite it.
- What you expected. Say what you wanted the code to do and how the output falls short.
- What you tried. List the changes you already made to the prompt or settings and what happened after each.
Search first. Many questions already have a thread, and replying there keeps the evidence together. One problem per thread, please.
Never post secrets
Do not post API keys, access tokens, passwords, private keys, connection strings, session cookies or private repository URLs, even ones you think are expired. Check your prompt, your output and any config file before you paste it, because assistants often echo keys back. Replace the value with a placeholder such as YOUR_API_KEY. If you post a secret by mistake, revoke it with the service that issued it first, then follow the steps on the contact page so we can remove it.
Cite your sources
If you state what a tool does, what a licence allows or what a project's rules say, link the documentation, licence text, policy page, paper or commit you are relying on. If the claim comes from your own testing, say which tool and version you used. Experience is welcome. It should not be presented as documentation.
Writing an answer
Answer the question in the first two sentences, then explain. Show the prompt line, the setting or the code change, not only a description of it. If the answer depends on the tool or its version, say which one, and say when you are not sure it applies elsewhere.
Not allowed
- Promotion. No posts whose main purpose is to advertise a product, a service, a paid course or a referral link. Mentioning a tool you actually used, in an answer to the question asked, is fine. Repeating it in every thread is not.
- Personal attacks. Disagree with the answer, not the person who wrote it. No insults, harassment or posts that target a person.
- Personal information. No names, emails or details that identify a private person, including in pasted logs and commit histories.
Moderation
Moderators may edit or remove posts that break these guidelines. That includes editing out a secret, trimming promotion or removing an off-topic thread. We try to leave a note when we edit a post and to explain a removal, but we may act first and explain afterwards when a post exposes a secret or personal information. Repeated breaches can lead to a suspended account. If you think a decision was wrong, write to editor@codelifter.com and tell us which post you mean.
Links
Links to documentation, licences, standards, papers and repositories are welcome. Links in posts are marked nofollow.