Gemini cli agent
Skill richfrem/agent-plugins-skills/plugins/cli-agents/skills/gemini-cli-agent
Gemini CLI sub-agent system for cost-efficient analysis using the `gemini` binary. Use when piping large contexts to cheaper Google Gemini models (gemini-3-flash-preview, gemini-3.1-pro-preview) for security audits, architecture reviews, or QA analysis. For frontier models (Gemini 3.5 Flash and above), use agy-cli-agent instead.From its SKILL.md
npx -y skills add richfrem/agent-plugins-skills --skill gemini-cli-agentAssembled from the repository path, not quoted from the project. Check it against their README if it does not work.
One thing to look at
- 5 stars5 stars. Stars are a popularity signal and not a quality one, but at this level it is likely that nobody has read this closely except its author, and you would be relying on your own review.
SKILL.md
3.7 KB, 868 tokens by cl100k_base, as published. Nobody here has run it
[!WARNING] Gemini CLI consumer access ended June 18, 2026. Free, Pro, and Ultra users no longer have access. Only enterprise Gemini Code Assist Standard/Enterprise licenses retain the
geminibinary. If you don't have one of those licenses, this skill will not work β useagy-cli-agentinstead, which is now the primary path for Gemini model access.
Identity: The Gemini Sub-Agent Dispatcher (Standard: gemini-3-flash-preview)
You dispatch specialized analysis tasks to Gemini CLI sub-agents using the gemini binary.
[!IMPORTANT] Default model: gemini-3-flash-preview (cost-efficient). For deep reasoning, use gemini-3.1-pro-preview. For frontier models (Gemini 3.5 Flash+), use
agy-cli-agentinstead.
Minimal Working Pattern
gemini -m gemini-3-flash-preview -p "$(cat agents/persona.md)
---SOURCE CODE---
$(cat target.py)
---INSTRUCTION---
Perform a full code review. Use severity levels: π΄ CRITICAL, π‘ MODERATE, π’ MINOR.
You are operating as an isolated sub-agent.
Do NOT use tools. Do NOT access filesystem." > review.md
Orchestration Pattern: run_agent.py
python ./scripts/run_agent.py <PERSONA_FILE> <INPUT_FILE> <OUTPUT_FILE> "<INSTRUCTION>" [MODEL_NAME]
Mandatory Health Check
Before any complex orchestration:
gemini --yolo -m gemini-3-flash-preview -p "hello"python ./scripts/run_agent.py agents/refactor-expert.md target.py ./heartbeat.md "Verify health"- Confirm
./heartbeat.mdis not empty.
Example
python ./scripts/run_agent.py agents/security-auditor.md target.py security.md \
"Find vulnerabilities. Use severity levels: π΄ CRITICAL, π‘ MODERATE, π’ MINOR."
Persona Registry (agents/)
| Persona | Use For |
|---|---|
security-auditor.md | Red team, vulnerability scanning, threat modeling |
refactor-expert.md | Optimizing code for readability, performance, and DRY |
architect-review.md | Assessing system design, modularity, and complexity |
Capability Boundary
Image Generation NOT supported
The gemini binary is text and code only. Image generation models require the Python google-genai SDK with a paid billing account.
Model Capacity
gemini-3.1-pro-preview may hit MODEL_CAPACITY_EXHAUSTED (429) under load β retry or fall back to gemini-2.5-pro.
CLI Best Practices
Avoid Shell Expansion for Large Contexts
$(cat ...) > 10KB can silently fail in background processes. Use run_agent.py which writes to a temp file.
Autonomous Orchestration (--yolo)
Pass --yolo to allow all tool calls to run without confirmation for headless operation.
Workspace Boundary (IDEClient Directory Mismatch)
Always invoke gemini from your active workspace directory. To operate in an external folder, pass a cd instruction in the prompt itself.
Backgrounding & TTY (SIGTTIN)
nohup gemini --yolo -m gemini-3-flash-preview -p "..." >> log.txt 2>&1 < /dev/null &
< /dev/null is critical to prevent SIGTTIN stops.
Update / Install
npm install -g @google/gemini-cli@latest
# or via npx (run_agent.py falls back to this automatically)
npx @google/gemini-cli
Smoke Test
gemini -p "hello"
python ./scripts/run_agent.py agents/refactor-expert.md target.py output.md "Refactor this code."
What ships with it: 15 files
8.8 KB alongside SKILL.md, 1 of them executable
evals/
- evals.json1.2 KB
- .lock.hashes583 B
- results.tsv389 B
- traces/iter_001_KEEP_score0.68.json4.6 KB
references/
scripts/
- run_agent.pyruns29 B
- acceptance-criteria.md808 B
- fallback-tree.md991 B
- requirements.txt22 B