v1.0.1

Byted Las Vlm Video

ReviewClawScan verdict for this skill. Analyzed May 1, 2026, 8:19 AM.

Analysis

The skill matches its video-analysis purpose, but it asks for cloud credentials and automatically installs an external SDK package during setup, so it should be reviewed before use.

GuidanceBefore installing, review the automatic SDK installation path, use a dedicated least-privilege LAS API key, confirm any upload and estimated cost before processing, avoid sending sensitive videos unless appropriate, and stop any background polling process when it is no longer needed.

Findings (4)

Artifact-based informational review of SKILL.md, metadata, install specs, static scan signals, and capability signals. ClawScan does not execute the skill or run runtime probes.

Abnormal behavior control

Checks for instructions or behavior that redirect the agent, misuse tools, execute unexpected code, cascade across systems, exploit user trust, or continue outside the intended task.

Agentic Supply Chain Vulnerabilities

SeverityMediumConfidenceHighStatusConcern

scripts/env_init.sh

manifest=$(curl -sf "$manifest_url" || echo '{}') ... pip install --quiet --upgrade https://las-ai-cn-beijing-online.tos-cn-beijing.volces.com/operator_cards_serving/public/skills/sdk/las_sdk-0.2.0-py3-none-any.whl

The environment initializer fetches a remote manifest and conditionally installs a remote wheel with pip, but the package contents, hash, and install spec are not provided in the artifacts.

User impactRunning the normal setup can execute newly downloaded SDK code in the user's environment before the user has reviewed that code.

RecommendationRequire explicit user approval for the install, pin the exact package version and hash, and document this dependency in the install metadata.

Rogue Agents

SeverityLowConfidenceHighStatusNote

scripts/poll_background.sh

# Poll until terminal state
while true; do ... sleep "$SLEEP"
done

The included polling helper can continue running until the remote task reaches a terminal state, with no explicit maximum runtime.

User impactIf invoked, the helper could remain active for a long time and continue making provider polling calls.

RecommendationRun background polling only when needed and add a maximum runtime or manual stop instructions.

Permission boundary

Checks whether tool use, credentials, dependencies, identity, account access, or inter-agent boundaries are broader than the stated purpose.

Identity and Privilege Abuse

SeverityLowConfidenceHighStatusNote

SKILL.md

- **环境变量**: `LAS_API_KEY` (必填) ... 需要 `VOLCENGINE_ACCESS_KEY` 和 `VOLCENGINE_SECRET_KEY`

The skill requires a LAS API key and may request Volcengine storage credentials for output downloads, although the registry metadata lists no required environment variables.

User impactThese credentials can authorize provider API usage, billing, and possibly cloud storage access.

RecommendationUse least-privileged, project-specific credentials; avoid pasting secrets into chat; and revoke or rotate keys after use if they were exposed.

Sensitive data protection

Checks for exposed credentials, poisoned memory or context, unclear communication boundaries, or sensitive data that could leave the user's control.

Insecure Inter-Agent Communication

SeverityLowConfidenceHighStatusNote

SKILL.md

如果用户要求处理的是**本地文件**，则需要先通过 File API 上传至 TOS ... lasutil file-upload <local_path>

Local video files are uploaded to Volcengine/TOS and then analyzed through the LAS VLM provider flow, which is expected for this skill but moves user content outside the local environment.

User impactPrivate or sensitive video content may be processed and stored by the external provider.

RecommendationConfirm the video is appropriate to upload, prefer least-exposed TOS paths or signed URLs, and review provider retention and access policies.