Is Easel safe?
- Python shell/command execution
- Node.js shell/command execution
- Defense-evasion command idiom
What to do: Read the findings below before installing: each one opens the exact line of code it was found on.
Easel is a PyPI package analyzed by SkillTotal's deterministic static scanner. The scan found no malicious indicators, though 2 risky constructs are reported for review. It can: filesystem read, filesystem write, install time execution, network egress and shell execution — capabilities are what the code can do, not a verdict on intent. Risk score 30/100 (medium).
easel 0.2.1
Automated static-analysis result. It can contain false positives and false negatives, and is not a claim about the intent of Easel's authors. Report a false positive.
Behavioral traits
How this component maps to the CSA agentic threat model. Descriptive — it never affects the risk score.
Findings (10)
package.json runs scripts automatically when the package is installed.
"postinstall": "node scripts/postinstall.js",
Why it matters: Install scripts are a favorite supply-chain foothold — they execute on every machine that installs the package.
Fix: Inspect the hook command. Install-time scripts are a common supply chain execution vector; ensure they do nothing beyond a documented build step.
A command uses a known defense-evasion idiom: PowerShell execution-policy bypass / encoded command / hidden window, macOS code-signing bypass, or launching a payload from a world-writable temp directory. These are hallmarks of droppers and rarely appear in legitimate code. (2 occurrence(s) shown as evidence).
-WorkingDirectory $Root -RedirectStandardOutput $LogFile -RedirectStandardError $ErrorLogFile -WindowStyle Hidden | Out-Null
& powershell -NoProfile -ExecutionPolicy Bypass -File (Join-Path $Root 'scripts\gateway.ps1') start
Fix: Verify why the component bypasses execution policy / code signing or runs from a temp directory; these patterns are characteristic of malware staging.
The component can run operating-system commands or spawn processes.
const result = Bun.spawnSync({import { spawn, execFileSync, type ChildProcess } from "node:child_process";const child = spawn(chromeBinary, args, {Why it matters: Powerful and often legitimate — confirm the commands aren't built from untrusted input.
Fix: Confirm the command and its arguments are fully controlled and not derived from untrusted input; prefer execFile with an argument array.
The component can run operating-system commands or spawn processes.
result = subprocess.run(cmd, cwd=str(PROJECT_ROOT), env=_proxy_env())
result = subprocess.run(
[sys.executable, str(script)],
cwd=str(PROJECT_ROOT),
env=env,
)result = subprocess.run(
["node", "--version"],
capture_output=True, text=True, timeout=10,
)result = subprocess.run(
openclaw_base_cmd() + ["--version"],
capture_output=True, text=True, timeout=10,
)result = subprocess.run(command, cwd=PROJECT_ROOT)
result = subprocess.run(
cmd,
capture_output=True, text=True, timeout=timeout,
env=env,
)result = subprocess.run(cmd, capture_output=True, text=True,
cwd=str(PROJECT_ROOT), timeout=timeout + 30,
env=_proxy_env())result = subprocess.run(
openclaw_base_cmd() + ["--version"],
capture_output=True, text=True, timeout=10,
)proc = subprocess.run(argv, capture_output=True, text=True, encoding="utf-8",
errors="replace", timeout=20)r = subprocess.run(cmd, capture_output=True, text=True)
r = subprocess.run(
["ffprobe", "-v", "error", "-show_entries", "format=duration",
"-of", "default=nw=1:nk=1", path],
capture_output=True, text=True)r = subprocess.run(
["ffprobe", "-v", "error", "-select_streams", "a:0",
"-show_entries", "stream=codec_name", "-of", "csv=p=0", path],
capture_output=True, text=True)r = _sp.run([sys.executable, __file__, "assemble",
"--storyboard", str(_wsb(wd, vid, 3.0)), "-o", str(wd / "x.mp4"),
"--pad-mode", "trim"], capture_output=True, text=True)r = subprocess.run(cmd)
cp = subprocess.run(["ffprobe", "-v", "error", "-select_streams", "a:0",
"-show_entries", "stream=codec_name", "-of", "csv=p=0", str(path)],
text=True, capture_output=True)cp = subprocess.run([sys.executable, str(asr_script), "transcribe", "-i", str(rp),
"--model", model, "--format", "json", "-o", str(asr_out)],
text=True, capture_output=True)v = subprocess.run(["biliup", "--version"], capture_output=True, text=True, env=_direct_env())
return subprocess.call(["biliup", "-u", cookie, "login"], env=_direct_env())
rc = subprocess.call(cmd, env=_direct_env())
return subprocess.run(cmd, env=env, check=False).returncode
proc = subprocess.run(
[cli_path, "--version"],
capture_output=True,
text=True,
timeout=10,
)proc = subprocess.run(
cmd,
capture_output=True,
text=True,
timeout=timeout,
env=os.environ.copy(),
)proc = _sp.run(cmd, capture_output=True, text=True, timeout=300)
r = subprocess.run(cmd, capture_output=True, text=True, encoding="utf-8", errors="replace")
out = subprocess.check_output(
["ffprobe", "-v", "error", "-show_entries",
"format=duration:stream=width,height,codec_type",
"-of", "json", str(p)],
stderr=subprocess.STDOUT).decode()Why it matters: Powerful and often legitimate — confirm the commands aren't built from untrusted input.
Fix: Confirm the command and its arguments are fully controlled and not derived from untrusted input; avoid shell=True.
A server is bound to all network interfaces (0.0.0.0), not just your own machine.
uvicorn.run(app, host="0.0.0.0", port=port, log_level="warning")
Why it matters: Without authentication, other hosts on the network can reach it.
Fix: Bind to 127.0.0.1 for local-only use, or require authentication and restrict access if remote exposure is intended.
The component reads files from disk.
const fileData = fs.readFileSync(filePath);
Why it matters: Usually legitimate, but worth confirming it can't be steered into reading sensitive files.
Fix: Confirm which files are read and that paths cannot be influenced by untrusted input to reach sensitive locations.
The component reads files from disk.
for line in identity.read_text(encoding="utf-8").splitlines():
for line in env_file.read_text(encoding="utf-8").splitlines():
cfg = json.loads(cfg_path.read_text(encoding="utf-8"))
return p.read_text(encoding="utf-8")
raw = (state_dir() / CONFIG_FILENAME).read_text(encoding="utf-8")
data = json.loads(DEVICE_IDENTITY_FILE.read_text(encoding="utf-8"))
data = json.loads(cfg.read_text(encoding="utf-8"))
data = json.loads(manifest.read_text(encoding="utf-8"))
for line in path.read_text(encoding="utf-8", errors="replace").splitlines():
cfg = json.loads(config_path().read_text(encoding="utf-8"))
text = filepath.read_text(encoding="utf-8").strip()
text = filepath.read_text(encoding="utf-8").strip()
with build_opener(ProxyHandler({})).open(request, timeout=600) as response:upstream = build_opener(ProxyHandler({})).open(request, timeout=600)return json.loads(mj.read_text(encoding="utf-8"))
existing = json.loads(mpath.read_text(encoding="utf-8"))
data = json.loads((proj / MANIFEST_NAME).read_text(encoding="utf-8"))
lines = open(path, encoding="utf-8")
with build_opener(ProxyHandler({})).open(request, timeout=600) as response:raw = path.read_text(encoding="utf-8").splitlines()
text = path.read_text(encoding="utf-8")
tree = ast.parse(script.read_text(encoding="utf-8"))
lines = path.read_text(encoding="utf-8").splitlines()
lines = path.read_text(encoding="utf-8").splitlines()
text = path.read_text(encoding="utf-8")
Why it matters: Usually legitimate, but worth confirming it can't be steered into reading sensitive files.
Fix: Confirm which files are read and that paths cannot be influenced by untrusted input to reach sensitive locations.
The component writes or deletes files on disk.
tmp.write_text(json.dumps(data, ensure_ascii=False, indent=2), encoding="utf-8")
shutil.move(str(src), str(dst))
(proj / "note.md").write_text("正文", encoding="utf-8")(proj / "cover.png").write_bytes(b"\x89PNG")
(proj / "card_1.png").write_bytes(b"\x89PNG")
(proj / "card_1.jpg").write_bytes(b"jpg") # 重复 jpg
(proj / "export_cards.py").write_text("build", encoding="utf-8")(proj / "frames" / "f001.png").write_bytes(b"\x89PNG")
(proj / "meta.json").write_text(json.dumps(
{"title": "真标题", "platform": "小红书", "tags": ["A", "B"]}, ensure_ascii=False),
encoding="utf-8")bak.write_text("\n".join(raw) + "\n", encoding="utf-8")Path(tmp).write_text(
"".join(json.dumps(r, ensure_ascii=False) + "\n" for r in clean)
+ ("".join(l + "\n" for l in bad_lines) if bad_lines else ""),
encoding="utf-8")p.write_text("\n".join(json.dumps(r, ensure_ascii=False) for r in rows) + "\n", encoding="utf-8")os.remove(tmp)
srt_path.write_text("\n".join(lines), encoding="utf-8")ass_path.write_text(header + "\n".join(events) + "\n", encoding="utf-8")
concat_list.write_text("".join(f"file '{c}'\n" for c in clips), encoding="utf-8")shutil.copy(str(stage), str(out))
sbf.write_text(json.dumps(sb), encoding="utf-8")
srt.write_text("1\n00:00:00,000 --> 00:00:01,500\n第一句\n\n"
"2\n00:00:01,500 --> 00:00:03,000\n第二句\n", encoding="utf-8")p.write_text(json.dumps({"size": "1080x1920",
"shots": [{"video": str(vid), "duration": d}]}), encoding="utf-8")sbx.write_text(json.dumps({"size": "1080x1920",
"shots": [{"image": imgs[0], "duration": 4.0, "caption": "音效测试"}],
"sfx": [{"file": str(sfx), "at": 2.0, "volume": 0.8}]}), encoding="utf-8")native_sb.write_text(json.dumps({"size": "320x568", "pad_mode": "trim",
"shots": [{"video": str(native_src), "duration": 1.0, "audio_mode": "native"}]}),
encoding="utf-8")native_mix_sb.write_text(json.dumps({"size": "320x568", "pad_mode": "trim",
"narration": str(voice),
"shots": [{"video": str(native_src), "duration": 1.0, "audio_mode": "native"}]}),
encoding="utf-8")dub_sb.write_text(json.dumps({"size": "320x568", "pad_mode": "trim",
"shots": [{"video": str(native_src), "duration": 1.0, "audio_mode": "dub"}]}),
encoding="utf-8")with open(OUT_ASS, "w") as f:
Why it matters: Usually legitimate, but worth confirming the paths can't be controlled by untrusted input.
Fix: Confirm which files are written/deleted and that paths cannot be influenced by untrusted input.
The component makes outbound network requests.
const res = await fetch(`${baseURL}/images/generations`, {const res = await fetch(`${baseURL}/chat/completions`, {const res = await fetch(url);
analyzeViralCmd.option("--comment-pages <n>", "Comment pages to fetch (max 10)", "3");const res = await fetch(`${CDP_BASE}/json/version`, {const res = await fetch(`${CDP_BASE}/json/version`, {const res = await fetch(url, {const res = await fetch(url, {const res = await fetch(`${host}${fullUri}`, {const res = await fetch(`${host}${uri}`, {const res = await fetch(url, {const res = await fetch(url, {const res = await fetch(url, {fetch('/api/upload/limits').then((r) => r.json())const res = await fetch(`${BASE}${url}`, { cache: 'no-store', ...options });const res = await fetch(`${BASE}/api/chat/question/answer`, {const res = await fetch(`${BASE}/api/chat/question/status`, {? await fetch(`${BASE}/api/chat/stream`, {: await fetch(`${BASE}/api/chat/jobs/${encodeURIComponent(turnId || '')}/stream?after=${lastEventId}`, {Why it matters: Usually legitimate, but confirm the destinations are expected and no sensitive data leaves.
Fix: Confirm the destination hosts are expected and that no sensitive data is sent off-host.
The component makes outbound network requests.
import urllib.request
import urllib.request
with urllib.request.urlopen(healthz_url(), timeout=10) as response:
from urllib.request import ProxyHandler, Request, build_opener
request = Request(
endpoint,
data=json.dumps(payload, ensure_ascii=False).encode(),
headers={"api-key": api_key, "Content-Type": "application/json"},
method="PO …with build_opener(ProxyHandler({})).open(request, timeout=600) as response:request = Request(
stream_endpoint_for(endpoint),
data=json.dumps(payload, ensure_ascii=False).encode(),
headers={"api-key": api_key, "Content-Type": "application/json"},
method="POST", …upstream = build_opener(ProxyHandler({})).open(request, timeout=600)from urllib.request import ProxyHandler, Request, build_opener
request = Request(
os.environ.get("OPENAI_MAAS_ENDPOINT", DEFAULT_ENDPOINT),
data=raw_body,
headers={
# Standard OpenAI-compatible authentication, configurable for …with build_opener(ProxyHandler({})).open(request, timeout=600) as response:import http.client
import urllib.request
if urllib.parse.urlsplit(base_url).hostname in {request = urllib.request.Request(
url, data=body,
headers={"Authorization": f"Bearer {api_key}", "Content-Type": "application/json", "User-Agent": UA},
method="POST",
)with urllib.request.urlopen(request, timeout=timeout) as response:
request = urllib.request.Request(
url, headers={"Authorization": f"Bearer {api_key}", "User-Agent": UA}, method="GET",
)with urllib.request.urlopen(request, timeout=timeout) as response:
dl_req = urllib.request.Request(image_url, headers={"User-Agent": UA})with urllib.request.urlopen(dl_req, timeout=120) as resp:
dl_req = urllib.request.Request(image_url, headers={"User-Agent": UA})with urllib.request.urlopen(dl_req, timeout=120) as resp:
path = urllib.parse.urlparse(url).path
import urllib.request
handlers = [urllib.request.ProxyHandler(proxies)] if proxies else [
Why it matters: Usually legitimate, but confirm the destinations are expected and no sensitive data leaves.
Fix: Confirm the destination hosts are expected and that no sensitive data is sent off-host.
How attackers abuse these capabilities
Interactive labs on the attack class behind the rules above. They show the technique, not anything found in Easel.
Check your own component
Run the same evidence-backed scan on any MCP server, agent skill, or package.
Scan your own componentHow we determine this: deterministic static analysis (regex + AST), evidence-anchored, no code execution. Methodology →