gemini-cli-agent — independently scanned and version-tracked by SaferSkills.
SaferSkills independently audited gemini-cli-agent (Agent Skill) and scored it 87/100 (green). The audit ran 55 deterministic rules across Security, Supply Chain, Maintenance, Transparency, and Community; it found 0 high-severity and 3 lower-severity findings. The full rule-by-rule trace and per-finding evidence are below. Free, methodology-open.
Findings & checks · 3 flagged
The text {match} tells the agent to skip the normal "ask the user first" gate. Used adversarially it removes the human-in-the-loop check before destructive or sensitive actions, turning a normally-gated agent into a fire-and-forget executor.
Every scanned point with the score it earned and what moved between them.
First recorded scan — no prior version to compare against.
The primary manifest — the file an agent reads to learn what this artifact does.
[!WARNING] Gemini CLI consumer access ends June 18, 2026. Free, Pro, and Ultra users lose access on that date. Only enterprise Gemini Code Assist Standard/Enterprise licenses retain thegeminibinary. The replacement for consumer users is the Antigravity (`agy`) CLI — useagy-cli-agentinstead.
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-agent instead.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: 🔴 CRITICAL, 🟡 MODERATE, 🟢 MINOR.
You are operating as an isolated sub-agent.
Do NOT use tools. Do NOT access filesystem." > review.mdrun_agent.pypython ./scripts/run_agent.py <PERSONA_FILE> <INPUT_FILE> <OUTPUT_FILE> "<INSTRUCTION>" [MODEL_NAME]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"./heartbeat.md is not empty.python ./scripts/run_agent.py agents/security-auditor.md target.py security.md \
"Find vulnerabilities. Use severity levels: 🔴 CRITICAL, 🟡 MODERATE, 🟢 MINOR."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 |
The gemini binary is text and code only. Image generation models require the Python google-genai SDK with a paid billing account.
gemini-3.1-pro-preview may hit MODEL_CAPACITY_EXHAUSTED (429) under load — retry or fall back to gemini-2.5-pro.
$(cat ...) > 10KB can silently fail in background processes. Use run_agent.py which writes to a temp file.
--yolo)Pass --yolo to allow all tool calls to run without confirmation for headless operation.
Always invoke gemini from your active workspace directory. To operate in an external folder, pass a cd instruction in the prompt itself.
nohup gemini --yolo -m gemini-3-flash-preview -p "..." >> log.txt 2>&1 < /dev/null &< /dev/null is critical to prevent SIGTTIN stops.
npm install -g @google/gemini-cli@latest
# or via npx (run_agent.py falls back to this automatically)
npx @google/gemini-cligemini -p "hello"
python ./scripts/run_agent.py agents/refactor-expert.md target.py output.md "Refactor this code."~30 seconds. Free. No account. Every finding cites a rule and a line of evidence.