Security audit
Security checks for vulnerabilities and agentic risk
Review before installing: potentially paid translations are treated as safe reads, and setup can execute an unverified remote installer.
Require authorization for the selected image, translation, and spending before execution; do not treat translation as a free read or retry blindly after an error. Review the CLI installer and its bundled-skill changes before approving setup. Installation and login support the integration, and no malicious behavior was established, but the external installer and hosted service were not audited.
The supplied command performs remote installer execution, not the declared ImageTranslate.AI search or read operations. Nothing in the supplied chunk establishes this installer as a supporting dependency for those operations. The downloaded script's contents and any further effects are not available for assessment.
The supplied chunk consists solely of a login command and shows no ImageTranslate.AI integration or data-access behavior. Authentication could support a matching implementation, but here it is the entire visible operation, making the apparent primary purpose materially different. The command alone does not establish which service is authenticated or whether credentials are persisted.
The curl-to-bash pipeline executes whatever the installation endpoint returns with the invoking user's privileges, without pinning or independent integrity verification. HTTPS reduces transport interception risk but does not protect against a compromised publisher or distribution endpoint; successful exploitation could expose local files and credentials. This overlaps with the setup-safety finding rather than representing a separate execution path.
oo: command not found — install the oo CLI (other platforms: https://cli.oomol.com/install-guide.md):
curl -fsSL https://cli.oomol.com/install.sh | bash # macOS / Linux
The metadata describes searching and reading, but the only exposed action submits an image translation that may incur charges and produces an output artifact. This mismatch can cause an agent to treat a billable operation as read-only, particularly because the safety section explicitly labels untagged actions as safe reads.
The translation action is untagged, so the skill's own policy permits running it directly as a safe read despite explicitly acknowledging that it can incur charges. A failed oversized-result check can still leave the user billed, making automatic execution or retries financially consequential.
The setup instructions direct immediate execution of downloaded shell or PowerShell code without an explicit approval checkpoint, integrity verification, or explanation of local changes. The missing-CLI condition limits when this occurs, and tool restrictions may require additional approval, but neither makes the downloaded code trustworthy if execution is permitted.
Skill enables autonomous high-impact decisions without human-in-the-loop verification. Critical operations (destructive commands, financial transactions, data deletion) should require explicit user confirmation.
- **`OOMOL_INSUFFICIENT_CREDIT`** — OOMOL billing stop. Recharge at `https://console.oomol.com/billing/token-recharge` before retrying.
- **HTTP 402 without `OOMOL_INSUFFICIENT_CREDIT`** — the third-party account itself has no active plan or credit. Report the provider error message as is; do not ask the user to recharge OOMOL.
## Resources