Caveat verdict

agent-security-skill-scanner

🟠 Risky

Significant risk patterns flagged — automated deep scan, not behavioral proof.

The skill scans for security issues but also executes code, which could potentially be exploited if not properly handled.

⚠ Flagged for review — coarse, uncorroborated signal, not a confirmed exploit. Review the config yourself before installing.

Automated static analysis — not a human review. Caveat flags capabilities, not confirmed intent, and can produce false positives. Disagree with this verdict? Use Dispute below.

security

transparency

100

maintenance

Permission integrity

Makes network requests but does not declare curl/wget in required binaries

network_out

Installs packages at runtime — transitive dependencies are not auditable

package_install

Findings (24)

Pattern match high

Uses eval() — can execute arbitrary code

config_detector.py · prose · downgraded · eval(

Pattern match high

Pipe-to-shell pattern (curl | sh) — supply chain attack vector

curl_risk_classifier.py · prose · downgraded · curl | bash

Pattern match high

Pipe to bash — executes piped content as shell commands

curl_risk_classifier.py · prose · downgraded · | bash

Pattern match high

Pipe to sh — executes piped content as shell commands

curl_risk_classifier.py · prose · downgraded · |sh

Pattern match high

Pipe to python — executes piped content as Python code

curl_risk_classifier.py · prose · downgraded · |python

Pattern match high

Bash /dev/tcp — raw TCP connection via shell

curl_risk_classifier.py · prose · downgraded · /dev/tcp/

Pattern match high

Accesses sensitive system files

src/engines/__init__.py · prose · downgraded · /etc/shadow

Pattern match high

Recursive delete from root or home — destructive command

src/engines/hybrid_scanner_v2.py · prose · downgraded · rm -rf /

Pattern match high

Accesses AWS credentials file

whitelist_filter.py · prose · downgraded · ~/.aws/credentials

Pattern match medium

Uses exec() — may execute shell commands

config_detector.py · prose · downgraded · exec(

Pattern match medium

Accesses .ssh directory

credential_theft_classifier.py · prose · downgraded · .ssh/

Pattern match medium

References sudo — requests elevated privileges

src/engines/__init__.py · prose · downgraded · sudo

Pattern match medium

setuid — privilege escalation mechanism

src/engines/__init__.py · prose · downgraded · setuid

Pattern match medium

References child_process — can spawn system processes

src/engines/pattern_engine.py · prose · downgraded · child_process

Pattern match medium

subprocess execution — runs system commands from Python

whitelist_filter.py · prose · downgraded · subprocess.call(

Pattern match medium

os.system/popen — direct OS command execution

whitelist_filter.py · prose · downgraded · os.system(

Pattern match medium

Raw socket connection — low-level network access

whitelist_filter.py · prose · downgraded · socket.connect(

Pattern match medium

Accesses cloud provider credentials

whitelist_filter.py · prose · downgraded · ~/.aws

Pattern match low

Python os.environ.get — reads environment variable

scanner.py · prose · downgraded · os.environ.get(

Pattern match low

Python urllib.request — network access

src/engines/hybrid_scanner_v2.py · prose · downgraded · urllib.request

Pattern match low

POSTs data to external URL

src/engines/llm_engine.py · prose · downgraded · .post('https://

Pattern match low

Base64 encoding/decoding

src/engines/pattern_engine.py · prose · downgraded · base64_decode

Pattern match low

Python shutil file operation — copies/moves/deletes files

whitelist_filter.py · prose · downgraded · shutil.rmtree(

Pattern match low

Python os.getenv — reads environment variable

whitelist_filter.py · prose · downgraded · os.getenv(

Why the tier is capped

Execution sink present in raw bytes (Hard Floor: class A/D/E). Final tier capped at Caution — cannot be lifted by any downgrade, example-payload opt-in, or allowlist.

Permissions & capabilities

No declared permissions — minimal attack surface.

package_installnetwork_out

Is this flag fair?

Check another skill Browse the registry Auditing your own skills or configs? Use the API