Skill flagged — suspicious patterns detected

ClawHub Security flagged this skill as suspicious. Review the scan results before using.

baidu-mapbox-isochrone

v1.0.3

生成等时圈(isochrone)。通过百度地理编码API将地址转为BD-09坐标,转换为WGS84后,调用Mapbox Isochrone API生成等时圈SHP文件和Python预览图。当用户需要:给定地址生成等时圈地图、百度坐标转WGS84后做等时圈分析、生成等时圈shapefile并提供预览图时使用此技能。

0· 29·0 current·0 all-time
MIT-0
Download zip
LicenseMIT-0 · Free to use, modify, and redistribute. No attribution required.
Security Scan
VirusTotalVirusTotal
Benign
View report →
OpenClawOpenClaw
Suspicious
medium confidence
!
Purpose & Capability
The skill claims to convert Baidu BD-09 → WGS84 and call Mapbox isochrone — that matches the included script. However, SKILL.md tells the agent to read Baidu AK and Mapbox AK from MEMORY.md but the skill declares no required env vars or credentials. The skill also mandates sending outputs via a 'feishu' message channel even though no Feishu credentials or channel setup are declared. These mismatches mean required capabilities/credentials are not documented in the metadata.
!
Instruction Scope
SKILL.md instructs the agent to run scripts, zip all generated artifacts and then automatically send a preview image and the ZIP package to a Feishu channel via the message tool (the doc marks this as 'must execute'). That is broader than simply producing files: it causes automatic outbound transfer of generated files (and potentially any other files placed in the output dir). The instructions also reference MEMORY.md as the source of API keys, giving the skill implicit access to stored secrets. The 'automatic send' requirement is an operational decision that may be undesirable/unexpected for users.
Install Mechanism
No install spec is provided (instruction-only with a bundled Python script). That minimizes additional install-time risk; the Python script will run when invoked and pulls remote tiles/APIs at runtime. No off-repo downloads or installers are specified in skill metadata.
!
Credentials
The skill needs two API keys (Baidu AK and Mapbox AK) according to SKILL.md, but it does not declare required environment variables or primary credentials. Instead it expects those keys in MEMORY.md (agent memory). This is an inconsistency: the skill requires secrets but does not document them as required env vars. Also the automatic sending to Feishu implies the agent must have messaging/channel credentials configured — again not declared. The lack of declared credentials makes it hard to reason about what the skill will access/send.
Persistence & Privilege
always is false and the skill is user-invocable (normal). There is no installer or configuration that attempts to persistently modify other skills or elevate privileges. Note: autonomous invocation is allowed by platform default — combined with the automatic-send instructions this increases the impact surface, but the skill itself does not request persistent 'always' inclusion.
What to consider before installing
This skill appears to do what it says (convert Baidu coords, call Mapbox, save shapefiles and preview images), but there are several red flags you should consider before installing or running it: - Secrets handling: SKILL.md expects Baidu and Mapbox API keys stored in MEMORY.md but the skill metadata does not declare required credentials. Confirm where these keys will be read from and that you are comfortable storing them there. - Automatic outbound sends: The instructions require zipping outputs and sending both the preview image and the ZIP to a Feishu channel via the message tool. If you do not expect files (or credentials) to be transmitted to that channel, do not run the skill until this behavior is removed or clarified. Ensure the message tool/channel is one you control and trust. - Verify produced bundle: The ZIP will include shapefiles and the TIFF. Make sure no other sensitive files are placed in the output directory (default /root/.openclaw/workspace/isochrone_output) before running. - Review the script: Although the visible portion looks legitimate (calls Baidu, Mapbox, and ESRI/Mapbox tiles), the shipped Python script should be fully reviewed (the snippet is truncated in the manifest) to ensure it does not exfiltrate extra data or read unexpected local files. If you lack the ability to review code, run it in an isolated sandbox with test API keys. - Documentation mismatch: Ask the author to declare required credentials in the skill metadata (and to make automatic sending optional). Prefer skills that declare needed env vars and do not auto-push outputs to external channels without explicit user consent. Given these inconsistencies (undocumented secrets + mandatory automatic sending), treat the skill as suspicious until the above concerns are addressed or you run it in a safe, controlled environment.

Like a lobster shell, security has layers — review code before you run it.

latestvk977zg8br9vbap8a5fsqz86h2184jejc

License

MIT-0
Free to use, modify, and redistribute. No attribution required.

Comments