Security audit
Security checks for vulnerabilities and agentic risk
This skill is a straightforward image-generation helper, but users should handle the API token carefully and understand prompts are sent to an external service.
Install only from a source you trust, consider pinning the installer or package version, and avoid putting real API tokens directly on the command line. Treat any prompt or reference image UUID you provide as data sent to api.talesofai.com, so do not use confidential business text or sensitive imagery unless that service's terms and retention policy are acceptable to you.
ogimage.js:10API Token Exposed Through Command-Line Arguments
SKILL.md:31Unpinned Third-Party Package Execution in Installation Instructions
The README promotes sending user prompts and optional reference image identifiers to a third-party image generation API but does not clearly warn that this data leaves the local environment. Users may unknowingly transmit sensitive prompts, identifiers, or business data to an external service, creating privacy, confidentiality, and compliance risks.
The README instructs users to install and run a remote package via npx skills without pinning an exact version. This creates a supply-chain risk because users may receive a newer or compromised package than the one the author originally tested, and installation-time or runtime code could execute with the user's privileges.
npx commands without a version suffix (e.g. @1.0.0) create a rug-pull risk if the upstream server is compromised and publishes a malicious update.
The skill declares only a Bash tool and documents execution of a Node script that sends prompts and an API token to an external Neta service, but it does not declare any explicit tool scope, permissions, or allowed network destinations. This creates an unnecessary trust gap: an agent or reviewer cannot easily determine or constrain outbound network behavior, increasing the risk of unintended data exfiltration or broader-than-expected network access.
The trigger guidance says to use the skill whenever someone asks to generate or create OG images, which is broad and lacks tighter constraints around when the skill should activate. Over-broad activation can cause unintended execution of shell/network-enabled functionality, exposing prompts or secrets to an external API when the user may only be asking a general question or comparing options.
The install instruction uses npx skills add SherriHidalgolt/og-image-skill without a pinned version or immutable reference. That means users may install changed content in the future, including malicious updates or a compromised package/repository state, undermining reproducibility and supply-chain integrity.
The skill accepts an authentication token and forwards it in an HTTP header to an external service without clear disclosure of where it is going. Combined with the provider mismatch, this creates a meaningful risk of credential exposure or misuse because users may believe they are authenticating to one vendor while the token is sent to another.
The skill claims to use the Neta API, but the code actually sends authentication material and user content to talesofai.com while presenting a different platform identity. This mismatch is dangerous because users may provide tokens and prompts under false assumptions about the recipient service, enabling credential misuse, undisclosed third-party data transfer, or deceptive routing of sensitive content.
The hardcoded outbound domain indicates data is sent to talesofai.com rather than the advertised provider. In a skill context, this increases risk because it can conceal where user data and credentials are actually being transmitted, which is a material trust and privacy issue rather than a harmless implementation detail.
// --- Submit job ---
async function makeImage() {
const res = await fetch("https://api.talesofai.com/v3/make_image", {
method: "POST",
headers: HEADERS,
body: JSON.stringify(body),
The hardcoded outbound domain indicates data is sent to talesofai.com rather than the advertised provider. In a skill context, this increases risk because it can conceal where user data and credentials are actually being transmitted, which is a material trust and privacy issue rather than a harmless implementation detail.
// --- Submit job ---
async function makeImage() {
const res = await fetch("https://api.talesofai.com/v3/make_image", {
method: "POST",
headers: HEADERS,
body: JSON.stringify(body),
The request sends user-supplied prompt text and optional reference identifiers to an external service without any user-facing disclosure in the skill itself. In this context, that is risky because prompts or reference UUIDs may contain sensitive or proprietary information, and the skill metadata already misstates the service being contacted.
The polling request transmits the auth header and task identifier to the same external service to retrieve job status. While status polling is expected behavior for asynchronous APIs, it still contributes to undisclosed third-party communication and credential use, which is more concerning because the service destination is inconsistent with the stated skill purpose.
for (let attempt = 0; attempt < maxAttempts; attempt++) {
await new Promise((r) => setTimeout(r, 2000));
const res = await fetch(`https://api.talesofai.com/v1/artifact/task/${taskUuid}`, {
headers: HEADERS,
});
The comment suggests a broader credential-resolution step, but the implementation strictly requires a user-supplied --token argument and exits otherwise. This is a mild intent/documentation mismatch because the documented section name implies token lookup or resolution behavior that is not actually present.
No suspicious patterns detected.