Skillv1.0.52

ClawScan security

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

Scanner verdict

SuspiciousMar 16, 2026, 3:50 AM

Verdict: suspicious
Confidence: medium
Model: gpt-5-mini
Summary: The skill's instructions match an image-generation use case, but there are modest inconsistencies (no declared credential requirements, no included code, and no source/homepage) that warrant caution before use.
Guidance: This skill looks like a straightforward image-generation helper, but exercise caution before installing or running anything: the SKILL.md expects you to provide an API key (example config shows 'api_key') and a local script (generate.py), yet the skill bundle does not include code or declare required credentials. Before using it: 1) verify the source/author (there's no homepage and owner ID is unknown); 2) request or inspect the generate.py implementation so you know what data is sent to external APIs; 3) never paste your long-lived API key into an unknown script—use a least-privilege or restricted key and monitor billing; 4) prefer official client libraries and endpoints (the doc references the 'openai' PyPI package and 'dall-e-3'); 5) if you plan to use this in production, ask the publisher for a homepage, privacy/data-use details, and a clear list of required environment variables. These steps will reduce the risk of accidental credential leakage or unexpected charges.

Review Dimensions

Purpose & Capability: noteThe skill describes text-to-image, image-variation, editing and batch generation which aligns with the examples in SKILL.md. However, the README sample expects an OpenAI-style API key and references the 'openai' package and model 'dall-e-3' while the skill metadata declares no required credentials or primaryEnv. That mismatch (documented API key usage vs no declared credential) is inconsistent.
Instruction Scope: noteSKILL.md only tells the user to pip install openai and pillow and run a local script (generate.py) and shows a JSON config with an api_key. The instructions do not ask the agent to read unrelated system files or exfiltrate data. However, no generate.py/implementation is included in the skill bundle, so the instructions assume the user/agent has or will supply executable code and an API key—this missing code is an operational gap and the doc references a secret (api_key) that isn't declared in metadata.
Install Mechanism: okThere is no install spec (instruction-only skill). The SKILL.md suggests 'pip install openai pillow' which is a normal PyPI install for this task. No arbitrary download URLs, extract steps, or nonstandard installers are present.
Credentials: concernThe documentation requires an 'api_key' in its example config and references using 'dall-e-3', but the skill metadata lists no required environment variables or primary credential. This is a proportionality/clarity issue: using the skill in practice will require supplying credentials (and likely billing), but the skill does not declare how or which credential names it expects. Also the skill source/homepage is unknown, so handing over API keys without verifying implementation origin increases risk.
Persistence & Privilege: okThe skill does not request persistent privileges (always is false), does not include installation hooks, and does not declare any config paths or system-level modifications. Nothing indicates it would alter other skills or system-wide agent settings.