Skillv1.0.0

ClawScan security

阿里云云效创建MR+发布+通知（全可配置） · ClawHub's context-aware review of the artifact, metadata, and declared behavior.

Scanner verdict

BenignApr 13, 2026, 8:04 AM

Verdict: benign
Confidence: high
Model: gpt-5-mini
Summary: The skill's code and runtime instructions match its stated purpose (operate on Aliyun 云效: create/query MR, trigger pipelines, send robot notifications) and the credentials it asks for are appropriate for that functionality.
Guidance: This skill appears internally consistent with its stated purpose. Before installing, consider: (1) Provide a least-privileged Yunxiao personal token (rotate it regularly) and ensure the token scope only allows repository/pipeline actions required. (2) Verify WECOM_WEBHOOK_URL points to your organization’s bot (do not use public/pasted webhook URLs). (3) Review the Python file yourself for any additional endpoints or unexpected behavior; the code makes HTTP calls to openapi-rdc.aliyuncs.com and to whatever webhook URL you supply. (4) Ensure the runtime has the 'requests' library available; the skill metadata does not declare dependencies. (5) Test with non-production repositories/pipelines first. (6) Avoid posting secrets into chat logs or public places — set them as environment variables on the host or in a secure secret store.

Review Dimensions

Purpose & Capability: okName/description, SKILL.md and the Python code all target Aliyun 云效 (openapi-rdc.aliyuncs.com) operations (repository lookup, create changeRequests, search/run pipelines) and sending notifications to a webhook. Required secrets (personal token, organization id, webhook) are exactly what you would expect for these operations.
Instruction Scope: okSKILL.md instructs the agent to parse natural language, call Yunxiao APIs and post results to the configured webhook. The instructions do not ask the agent to read unrelated system files or transmit data to unexpected third-party endpoints beyond the Yunxiao API domain and the user-supplied webhook URL.
Install Mechanism: okThere is no install spec and the skill comes with a single Python file; nothing is downloaded from external URLs and no archive extraction is used. This lowers install-time risk. Note: the code uses the 'requests' library but the skill metadata doesn't list runtime dependencies.
Credentials: noteThe sensitive environment variables documented in SKILL.md (YUNXIAO_PERSONAL_TOKEN, YUNXIAO_ORGANIZATION_ID, WECOM_WEBHOOK_URL) are appropriate for the described functionality. Minor inconsistencies: the registry metadata reported 'required env vars: none' while SKILL.md and the code clearly expect those environment variables (they are optional but recommended). Also the code relies on the 'requests' package which is not declared in metadata.
Persistence & Privilege: okThe skill does not request always:true, does not modify other skills or system-wide configs, and does not persist credentials itself. It will, however, send network requests to the Yunxiao API and to the configured webhook when invoked.