Claude has friends. On the calls that matter, it asks them — AIs from other families who think differently and catch the blind spots Claude can't see from the inside. Don't ship alone. Ask your friends.
 |
 |
 |
 |
 |
 |
/plugin marketplace add awrshift/claude-friends
Then set your Gemini key (free tier at aistudio.google.com):
pip install google-genai
# Store the key once — read automatically by every shell, no `export` needed:
mkdir -p ~/.gemini && printf '%s' 'your_key_here' > ~/.gemini/api_key && chmod 600 ~/.gemini/api_keyPrefer an env var?
export GOOGLE_API_KEY=your_key_hereworks too and overrides the file. A bareGOOGLE_API_KEY=...line in~/.envwill not work — nothing sources it in a fresh shell.
Each friend thinks in a different way, so they don't share the same blind spots:
| Friend | Who | Why they help |
|---|---|---|
| 🔷 Gemini | A friend from another family (Google) | Different training, different instincts — catches what Anthropic-family models miss |
| 🧠 A second Claude, fresh eyes | Same family, no memory of your chat | Can't get anchored on the framing you already talked yourself into |
| ⚫ GPT (optional) | A third family (OpenAI) | Widens the room for the highest-stakes calls — opt-in, needs a paid key |
| Ask… | Trigger phrase | What happens |
|---|---|---|
| A friend | "sanity check", "am I missing something" | One friend, quick gut-check (Gemini or the fresh-eyes Claude). ~3¢ |
| The whole table | "this is important", "gather the friends" | All friends at once, in parallel — for the calls that matter. ~7¢ |
| Hash it out | "help me choose", "brainstorm options" | A few rounds back and forth to converge on one path. ~25¢ |
Claude reads your intent and picks the right mode. Some natural ways to trigger each:
Want a quick gut-check from one friend?
"ask a friend" · "sanity check this" · "am I missing something" · "stress-test this" · "critique this" · "give me a second opinion" · "cross-check this" · "review this" · "thoughts?" · "is this right?" · "devil's advocate" · "poke holes in this" · "what could go wrong"
High-stakes — want the whole table?
"this is important" · "run a full review" · "gather the friends" · "ask everyone" · "check before I send" · "before publishing" · "big decision" · "high-stakes review" · "boardroom debate" · "two independent opinions" · "don't let me ship something dumb" · "this can't be wrong"
Choosing between options?
"help me choose between" · "brainstorm options" · "I have several paths" · "I have 3 angles on this" · "multiple options to weigh" · "diverge and converge" · "multi-round brainstorm" · "round-table discussion" · "compare these approaches" · "what's the best direction"
Calling one friend by name?
"ask Gemini" · "ask Opus" · "ask GPT" · "what would Gemini say" · "let's get a second model on this" · "another perspective please"
If none of these fit, just describe what you're stuck on. Claude figures out which friend to bring in.
SKILL.md — how it works, when to invoke, anti-patterns, CLI reference. Want a third family for the biggest calls? references/third-family-gpt.md adds GPT (opt-in).
This is the same skill, re-framed: it was never about not asking — it's about asking the right friends. The old repo slug (awrshift/skill-dont-ask-me) still resolves — GitHub redirects it to awrshift/claude-friends. If you had it installed as dont-ask-me, re-add the marketplace and install claude-friends; the CLI commands (gemini.py, gpt.py) are unchanged.
MIT. PRs welcome.