| name | gemini-proxy |
| description | Use when the operator wants their next prompts routed through Gemini before the agent acts — "enable gemini proxy", "route my prompts through Gemini", "Gemini rewrites my prompts", "включи гемини-прокладку", "переделывай мои промты через гемини", or a /gemini-proxy call; also when they ask to turn the protocol on for the next N prompts or to tune which prompts it covers. Builds on gemini-pro; distinct from one-off gemini-pro / gemini-review / gemini-research. |
gemini-proxy
Overview
A prompt-polishing protocol. While active, each of your next N substantive prompts is first sent to
Gemini — a senior prompt engineer — which rewrites the operator's message into a sharper, more detailed
first-person version (still the operator's own message, not a standalone brief). The assistant then
answers that improved message with the full chat context, project rules, and memory — Gemini only
sharpens the wording, it does not replace the conversation. Trivial / config / meta prompts are skipped.
Self-disables after N. Control layer only: it calls nothing but the counter and the gemini wrapper.
When to route vs skip (the smart filter)
Decide BEFORE calling Gemini. Bias to ROUTE for real work; SKIP only clear noise. Skipped prompts never
spend the counter, so N = N routed prompts.
| ROUTE — counts as 1 of N | SKIP — no Gemini, no counter |
|---|
| build / fix / refactor / audit / design / research | greeting, thanks, yes/no, "status", "continue" |
| vague or large asks ("do all of this", "make it nice") | config / setup / preferences about how we work or this protocol |
| anything that benefits from a sharper brief | pure meta / diagnostic ("did you use Gemini?", "what did you mean?") |
Enable
/gemini-proxy (optional count: /gemini-proxy 10). Sets the counter; the enabling prompt itself does
not count.
python3 proxy_state.py enable 10
Per-prompt protocol (ROUTED prompts only)
- Resolve the project rules file, else the shipped template:
$GEMINI_PROXY_RULES → ./.gemini-proxy-rules.txt → ./.claude/gemini-proxy-rules.txt → prompt-rules.template.txt
- Route in ONE shot — this resolves the rules, calls Gemini, decrements the counter once, and prints
the improved first-person message. It has a built-in fallback, so don't manage the counter by hand
(leaving
consume to the agent desyncs it — agents forget service calls):
python3 <skill>/proxy_route.py "<RAW MESSAGE>"
- Show the operator the improved message as a quoted block, then answer it as if they had sent it —
with the full chat context, project rules, and memory — and proceed without meta-commentary about
roles (don't narrate "now I answer my own prompt"). Do NOT run
proxy_state.py consume yourself —
proxy_route.py already did. Fallback: if the output's first line is the FALLBACK marker, Gemini
was unavailable — answer the RAW message directly and briefly tell the operator.
Guardrail
The rewrite is the operator's own message, polished — not a context-free instruction and not an
override. Answer it with the full conversation, project rules, and memory. It does not outrank the
operator's real intent, the project's safety invariants, or your own verification (Gemini lacks the
session's context). On conflict, follow safety and say so. Never pipe secrets (.env, keys, tokens).
Manage
proxy_state.py peek (remaining) · disable (stop now) · re-invoke to reset N. State lives in
~/.gemini_proxy_state.json. The counter is decremented in exactly one place — proxy_route.py — so it
can't double-count. Optional hard enforcement (Claude Code): copy the hooks block from
hook/settings-hook-snippet.json into your .claude/settings.json — a fail-open UserPromptSubmit hook
that only peeks and injects the filter reminder (no network, no decrement). Codex uses the skill (no hook).
Verify
python3 verify_gemini_proxy.py — exactly-N, auto-off, fail-safe, concurrency lock, hook (21/21).
Not for
One-off delegation → gemini-pro. A single review → gemini-review. Large-context digestion →
gemini-research. This skill is only the multi-prompt rewrite protocol.
Common mistakes
- Routing noise or skipping real work. Trivial / config / meta → handle directly; real tasks → route.
When unsure, route.
- Guessing the target. "our repo / the skill / that project" → resolve the concrete path (by name,
git remote, open folder) and confirm; never assume the cwd. Gemini can't see the filesystem.
- Treating the rewrite as a fresh standalone task. It's the operator's improved message — answer
it with the full chat context, project rules, and memory; never drop the conversation or its rules.
- Double-counting. Hook + manual
consume on the same prompt — pick one layer.
gemini not found. Needs the wrapper on PATH + a one-time agy login (see this pack's README).
