Skillv1.0.0

ClawScan security

Azure Image Gen · ClawHub's context-aware review of the artifact, metadata, and declared behavior.

Scanner verdict

SuspiciousFeb 16, 2026, 9:13 AM

Verdict: suspicious
Confidence: medium
Model: gpt-5-mini
Summary: The skill's code and README consistently implement an Azure DALL‑E image generator, but the registry metadata omits the required Azure credentials and the instructions contain user-specific absolute paths — these inconsistencies warrant caution before installing.
Guidance: This package appears to do what it claims (generate images via Azure OpenAI), but there are a few red flags to consider before installing: - Metadata mismatch: The registry entry lists no required environment variables, but both SKILL.md and the script require AZURE_OPENAI_ENDPOINT, AZURE_OPENAI_API_KEY, and AZURE_OPENAI_DALLE_DEPLOYMENT. Treat the script as needing these secrets. - Secrets handling: The script reads a .env in the skill directory and also reads environment variables. Avoid storing long‑lived or highly privileged credentials in plain .env files in shared locations. Use a least-privilege API key and rotate/revoke it if used for testing. - Paths in README: Example commands use an absolute path (/Users/abhi/...). Update examples to your local path before running. - Network behavior: The script will send your prompt and API key to your configured Azure OpenAI endpoint; ensure you trust that Azure resource and its billing/usage implications. - Code review: The script is short and readable; if you are unsure, run it in a sandbox or inspect/execute it in a controlled environment first. If you proceed: provide a dedicated Azure OpenAI key with minimal permissions, avoid committing the .env to source control, run the script in a non-privileged directory, and verify the endpoint and deployment names are correct. If uncertain about the source (homepage unknown, owner unverified), prefer testing in an isolated environment or request the publisher to correct the registry metadata to declare the required credentials.

Review Dimensions

Purpose & Capability: noteThe name/description, SKILL.md, and scripts/generate.py all describe an Azure OpenAI DALL‑E image generator — the requested network calls, outputs, and files are coherent with that purpose. However, the registry metadata claims no required environment variables or primary credential, while both the SKILL.md and the script require AZURE_OPENAI_ENDPOINT, AZURE_OPENAI_API_KEY, and AZURE_OPENAI_DALLE_DEPLOYMENT. This metadata omission is inconsistent and could mislead users.
Instruction Scope: noteSKILL.md instructs running the included Python script and only references the Azure OpenAI API and local output files (images, manifest.json, index.html). The script reads a .env located in the skill directory (Path(__file__).parent.parent / '.env') and otherwise only uses environment variables and the network to the Azure endpoint. Minor concern: SKILL.md examples use an absolute, user-specific path (/Users/abhi/...), which is non-portable and suggests the README was copied from a local environment without sanitization.
Install Mechanism: okNo install spec is provided and there are no downloads or extract steps — this is instruction-only with a single Python script. That reduces install-time risk; nothing in the package will be automatically fetched or executed beyond the shipped script.
Credentials: concernThe script requires three Azure-related environment variables (endpoint, API key, deployment name) which are appropriate for the stated purpose. However, the skill metadata does not declare these requirements nor a primary credential, creating a misleading representation of required secrets. Because the skill needs an API key (sensitive secret) but the registry metadata omits this, users may not realize they must supply a credential or understand the blast radius.
Persistence & Privilege: okThe skill does not request permanent inclusion (always: false) and does not modify other skills or system-wide settings. It writes generated images and a manifest/index.html to a user-specified output directory only, which is expected behavior for this functionality.