v1.1.0

VecML AutoML

BenignClawScan verdict for this skill. Analyzed May 1, 2026, 8:20 AM.

Analysis

The skill appears coherent and purpose-aligned, but it uploads CSV data to VecML and uses a VecML API key, so users should only use it with data they are comfortable sending to that service.

GuidanceBefore installing or running, make sure the CSV can be sent to VecML, confirm the target column and destination endpoint, and use a revocable API key. Expect the skill to create remote projects/models and to save prediction outputs locally.

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.

Tool Misuse and Exploitation

SeverityLowConfidenceHighStatusNote

vecml-pipeline.py

post("create_project", { ... "project_name": project ... }) ... post("train_automl_model", { ... "model_name": model_name ... })

The helper performs remote API actions that create projects and train models. These actions are purpose-aligned, but they mutate the user's remote VecML workspace.

User impactRunning the command can create persistent projects, collections, and models in the user's VecML account.

RecommendationConfirm project, collection, and model names before running, and clean up remote artifacts that are no longer needed.

Agentic Supply Chain Vulnerabilities

SeverityInfoConfidenceMediumStatusNote

metadata

Source: unknown ... No install spec — this is an instruction-only skill.

The package includes a Python helper but has limited source/provenance information and no install spec. No suspicious install behavior is shown, but provenance is worth noting.

User impactUsers have less provenance context for the bundled helper script than they would with a clearly linked source repository or pinned release.

RecommendationReview the bundled script and provider documentation before using it with sensitive data or production credentials.

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

vecml-pipeline.py

KEY = os.environ.get("VECML_API_KEY", "") ... "user_api_key": KEY

The skill requires a VecML API key and includes it in API requests. This is expected for a VecML integration, but it gives the script authority to act in the user's VecML account.

User impactAnyone running the skill with a valid key allows it to create projects, upload data, train models, and query account resources supported by the API.

RecommendationUse a scoped or revocable API key if possible, keep it out of logs and shared shell history, and rotate it if 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

vecml-pipeline.py

API = os.environ.get("VECML_API_URL", "https://aidb.vecml.com/api") ... "X": x_b64 ... "file_data": y_b64

The script sends encoded feature and label CSV data to an external API endpoint. This matches the AutoML purpose, but it means user data leaves the local environment.

User impactCSV contents may be uploaded to VecML for processing and model training.

RecommendationUse only with data that may be shared with VecML, and confirm the upload destination before running, especially if the VECML_API_URL environment variable is set.