Skillv1.0.2

ClawScan security

Ezviz Open PTZ Control · ClawHub's context-aware review of the artifact, metadata, and declared behavior.

Scanner verdict

BenignMar 19, 2026, 3:18 AM

Verdict: Benign
Confidence: high
Model: gpt-5-mini
Summary: The skill's code, instructions, and environment requirements are consistent with its stated purpose (controlling Ezviz PTZ devices); the main notable risk is the optional global token cache which shares tokens between processes running as the same user.
Guidance: This skill appears to do what it says, but review and follow these precautions before installing: - Use dedicated Ezviz AppKey/AppSecret with the minimal permissions required (do not use account master credentials). - If you do not want tokens persisted, set EZVIZ_TOKEN_CACHE=0 to disable caching; otherwise cached tokens are stored under /tmp/ezviz_global_token_cache/global_token_cache.json with 0600 permissions (usable by any process running as the same OS user). - Audit the included Python files (scripts/main.py and lib/token_manager.py) before running; the code performs network requests to https://openai.ys7.com (the declared Ezviz Open API). Ensure this domain is the expected provider for your region. - Run the skill under a dedicated system user or environment if you are concerned about other local processes reusing tokens. - Validate the code on your platform (some listed excerpts were truncated in the bundle preview — verify the full scripts run without syntax/name errors) and run in a sandbox or test account first. If you need higher assurance, request the full, untruncated source for inspection or run the tool in an isolated environment with test credentials.

Review Dimensions

Purpose & Capability: okName/description, required env vars (EZVIZ_APP_KEY, EZVIZ_APP_SECRET), and the code all align: the package calls Ezviz API endpoints to list devices, control PTZ, and manage presets. No unrelated services or credentials are requested.
Instruction Scope: okSKILL.md instructs only to provide Ezviz credentials, install requests, and optionally disable token caching. Runtime instructions and CLI examples call the documented Ezviz API endpoints (openai.ys7.com). The instructions do reference reading/writing a token cache in /tmp, which is relevant to token management and documented.
Install Mechanism: okThis is instruction-only with local Python scripts; dependencies are limited to the requests library. No external download or obscure install sources are used in the provided files.
Credentials: okOnly Ezviz credentials are required (EZVIZ_APP_KEY, EZVIZ_APP_SECRET) and are justified by the interaction with the Ezviz API. The SKILL.md explicitly recommends minimal-permission credentials. No unrelated secrets or system credentials are requested.
Persistence & Privilege: noteThe skill creates/uses a global token cache in the system temp directory (/tmp/ezviz_global_token_cache/global_token_cache.json) with file mode 0600. This is coherent for token reuse, but it means any other code running under the same OS user can also read/use the cached tokens. The skill is not marked always:true and does not modify other skills' configurations.