Caveat verdict

dev-factory

45
🟠 Risky
Significant risk patterns flagged — automated deep scan, not behavioral proof.

This skill autonomously discovers CVE vulnerabilities, generates security tools based on them via a 7-agent pipeline, and automatically creates and publishes GitHub repositories — the combination of sourcing from vulnerability databases and auto-deploying generated code without explicit per-run human approval is a genuine concern beyond the skill's broad capability set.

⚠ 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.

0
security
80
transparency
90
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 (10)

Pattern match high

subprocess execution — runs system commands from Python

ACP_TEST_PLAN.md · code · subprocess.run(

Pattern match high

Possible hardcoded credential

builder/integration/notion_sync.py · prose · downgraded · API_KEY='): return line.split(

Pattern match medium

Python urllib.request — network access

IMPROVEMENT_ANALYSIS.md · code · urllib.request

Pattern match medium

Python os.getenv — reads environment variable

discovery-integration.md · code · os.getenv(

Pattern match medium

subprocess with shell=True — command injection vector

builder/symphony/glm5_agent.py · prose · downgraded · subprocess.run( command, shell=True

Pattern match medium

References agent memory files

COMPLETE_TEST_REPORT.md · code · MEMORY.md

Pattern match medium

pip3 install — installs Python packages at runtime

TEST_RESULTS.md · code · pip3 install

Pattern match low

Python os.environ.get — reads environment variable

builder/correction/fixer.py · prose · downgraded · os.environ.get(

Pattern match low

Python shutil file operation — copies/moves/deletes files

builder/pipeline.py · prose · downgraded · shutil.rmtree(

Pattern match low

Popular HTTP library — network access

tests/unit/test_notion_sync.py · prose · downgraded · got

Why the tier is capped

Execution sink present in raw bytes (Hard Floor: class D). 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_innetwork_out

Is this flag fair?

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