[Feature Request] Claude should be aware of his capabilities

[adriangalilea]

Let me just quote myself from https://github.com/anthropics/claude-code/issues/1340

I remember something about --continue but at this pace I never know if I actually heard it or misremembering.

Also asked Claude for the feature, he assured me no such feature exist and pointed me to various issues with feature requests.

Now here I realize that it is indeed existing, so the actual problem is that Claude should RTFM, and probably have it as a tool call.

There's likely more events such as this one.

[hesreallyhim]

I doubt this will get much traction, I can't tell you the reasoning from a business POV but it literally says in the system prompt that it's not allowed to visit the doc site.

[hesreallyhim]

There are no other Anthropic products. Claude can provide the information here if asked, but does not know any other details about Claude models, or Anthropic’s products. Claude does not offer instructions about how to use the web application or Claude Code. If the person asks about anything not explicitly mentioned here, Claude should encourage the person to check the Anthropic website for more information.

If the person asks Claude about how many messages they can send, costs of Claude, how to perform actions within the application, or other product questions related to Claude or Anthropic, Claude should tell them it doesn’t know, and point them to ‘https://support.anthropic.com/’.

If the person asks Claude about the Anthropic API, Claude should point them to ‘https://docs.anthropic.com/’.

[adriangalilea]

I doubt this will get much traction, I can't tell you the reasoning from a business POV but it literally says in the system prompt that it's not allowed to visit the doc site.

Just putting it out there, I'm not particularly worried about this.

I also understand that context is scarce, and any new built-in feature is potentially impacting the system as a whole, but I do believe self-knowledge should be priority, who best to Claude itself to tell you what Claude can do? This simply has to be improved one way or another.

[hesreallyhim]

Just putting it out there, I'm not particularly worried about this.

Well, that's fine, but if it's a feature request, that's kinda like saying you're not too worried if it gets adopted, but you're making your opinion heard, so I get that, and I can definitely relate, but it seems like it's obviously something they thought about and deliberately blocked. I remember with chatGPT i would get better advice on using the OpenAI API from Claude than ChatGPT, it kept using the old endpoint to get completions, so I don't really know what goes into these decisions.

Maybe there's some risk factor they discovered if a "smart" LLM can "learn about itself."

[github-actions[bot]]

This issue has been inactive for 30 days. If the issue is still occurring, please comment to let us know. Otherwise, this issue will be automatically closed in 30 days for housekeeping purposes.

[github-actions[bot]]

This issue has been automatically closed due to 60 days of inactivity. If you're still experiencing this issue, please open a new issue with updated information.

[github-actions[bot]]

This issue has been automatically locked since it was closed and has not had any activity for 7 days. If you're experiencing a similar issue, please file a new issue and reference this one if it's relevant.