agy-cli-agent — independently scanned and version-tracked by SaferSkills.
SaferSkills independently audited agy-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.
You dispatch tasks to Google Gemini frontier models via the agy binary.
[!IMPORTANT]agyis the Antigravity CLI for frontier models only (Gemini 3.5 Flash and above). It is a separate binary fromgemini. For cost-efficient older models, usegemini-cli-agent.
| Use case | CLI | Binary |
|---|---|---|
| Frontier models (3.5 Flash+) | Antigravity | agy |
| Cheaper/older models (3-flash, 2.5-pro) | Gemini CLI | gemini |
agy --dangerously-skip-permissions -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>" --cli agyrun_agent.py calls agy --dangerously-skip-permissions -p and streams output live to stdout and the output file simultaneously.
agy -p "HEARTBEAT CHECK: Respond with HEARTBEAT_OK only."
# or via run_agent.py:
python ./scripts/run_agent.py /dev/null /dev/null ./heartbeat.md "HEARTBEAT CHECK: Respond HEARTBEAT_OK only." --cli agy
grep -q "HEARTBEAT_OK" ./heartbeat.md && echo "OK" || echo "FAIL"python ./scripts/run_agent.py agents/security-auditor.md target.py security.md \
"Find vulnerabilities. Use severity levels: 🔴 CRITICAL, 🟡 MODERATE, 🟢 MINOR." --cli agyagents/)| 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 |
| Flag | Purpose |
|---|---|
-p "prompt" / --prompt "prompt" | Pass prompt non-interactively |
--dangerously-skip-permissions | Headless mode — skip all permission prompts |
--sandbox | Run in sandboxed environment |
agy is typically at /opt/homebrew/bin/agy. Confirm with which agy.
$(cat ...) > 10KB can silently fail. run_agent.py writes to a temp file automatically.
nohup agy --dangerously-skip-permissions -p "..." >> log.txt 2>&1 < /dev/null &< /dev/null is required to prevent SIGTTIN stops in background processes.
agy -p "hello"
python ./scripts/run_agent.py agents/refactor-expert.md target.py output.md "Refactor this code." --cli agy~30 seconds. Free. No account. Every finding cites a rule and a line of evidence.