Security audit
Security checks across malware telemetry and agentic risk
The skill is mostly an audit and recommendation helper, but it also scans sensitive session histories and one helper script runs local diagnostic commands beyond the read-only description.
Install only if you are comfortable with a skill auditing local skills, plugins, project files, MCP config, and session history. Before running extract_usage_signals.py, scope the session directories and consider removing or disabling the headroom live probe unless you trust your PATH and local headroom binary. Treat lifecycle/action outputs as plans that require your explicit confirmation, backups, and review before any agent changes files or configuration.
pass
# 2. Find headroom binary
which_result = subprocess.run(
["which", "headroom"] if sys.platform != "win32" else ["where", "headroom"],
capture_output=True, text=True, timeout=10,
)# 3. Run headroom perf
if headroom_bin:
perf_result = subprocess.run(
[headroom_bin, "perf"], capture_output=True, text=True, timeout=30,
)
if perf_result.returncode == 0:process_found = False
try:
if sys.platform == "win32":
proc = subprocess.run(
["tasklist", "/FI", "IMAGENAME eq python.exe", "/V"],
capture_output=True, text=True, timeout=10,
)if "headroom" in proc.stdout.lower():
process_found = True
else:
proc = subprocess.run(
["ps", "aux"],
capture_output=True, text=True, timeout=10,
)VirusTotal findings are pending for this skill version.
Detected: suspicious.dynamic_code_execution