Caveat verdict
gpt-image2-ppt
The skill generates PPT slides by calling the official OpenAI Images API with gpt-image-2, rendering templates locally via LibreOffice or Docker, with all operations transparently documented and no unexpected data exfiltration.
⚠ 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.
Permission integrity
credential_access
agent_memory
Findings (9)
Recursive delete from root or home — destructive command
docs/install.md · code · rm -rf /
Pipe to bash — executes piped content as shell commands
docs/install.md · prose · downgraded · | bash
Possible hardcoded credential
install_as_skill.sh · prose · downgraded · API_KEY:"
print_info
subprocess execution — runs system commands from Python
scripts/codex_backend.py · prose · downgraded · subprocess.run(
References sudo — requests elevated privileges
scripts/render_template.py · prose · downgraded · sudo
apt-get install — installs system packages
scripts/render_template.py · prose · downgraded · apt-get install
Python os.getenv — reads environment variable
generate_ppt.py · prose · downgraded · os.getenv(
pip3 install — installs Python packages at runtime
install_as_skill.sh · prose · downgraded · pip3 install
Python os.environ.get — reads environment variable
scripts/render_template.py · prose · downgraded · os.environ.get(
Why the tier is capped
Execution sink present in raw bytes (Hard Floor: class A/B/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.
credential_accessnetwork_indir_traversalagent_memory Is this flag fair?
Thanks — recorded.