Cost Prediction
v2.0.0Predict construction project costs using Machine Learning. Use Linear Regression, K-Nearest Neighbors, and Random Forest models on historical project data. Train, evaluate, and deploy cost prediction models.
⭐ 0· 1.3k·6 current·6 all-time
MIT-0
Download zip
LicenseMIT-0 · Free to use, modify, and redistribute. No attribution required.
Security Scan
OpenClaw
Benign
high confidencePurpose & Capability
Name and description (train/evaluate/deploy cost-prediction models) match the SKILL.md and instructions.md. The declared filesystem permission aligns with reading historical CSVs and saving models. The only mild ambiguity is the word 'deploy' — the instructions constrain computation to local filesystem (no external APIs), so 'deploy' appears to mean saving models locally rather than deploying to a remote service.
Instruction Scope
Instructions focus on preparing local datasets, feature engineering, training LinearRegression/KNN/RandomForest, evaluating metrics, and saving models. They do not instruct reading unrelated system files, accessing environment secrets, or contacting external endpoints. The guidance to 'gather project parameters' means soliciting user input, not scanning system data.
Install Mechanism
There is no install spec (instruction-only), which is low risk because nothing is written or downloaded by the skill itself. However, runtime Python libraries (pandas, scikit-learn, numpy) are used in the examples but not declared; users must ensure those dependencies are present. This is a minor coherence/usability gap but not a security red flag.
Credentials
The skill requests no environment variables or external credentials. The filesystem permission declared in claw.json is proportionate to the stated need to read historical datasets and save trained models. Note: filesystem permission inherently allows reading any local files the agent process can access, so users should be aware of privacy of local data used for training.
Persistence & Privilege
always:false and normal autonomous-invocation settings are appropriate. The skill does not request to modify other skills or system-wide settings and has no install step that would create persistent background components.
Assessment
This skill appears coherent and local-only. Before using it, confirm you trust the CSVs or other local data you will load (they may contain sensitive project or personnel information). Ensure the Python environment has pandas, numpy, and scikit-learn installed (the SKILL.md examples assume these). If you prefer stricter isolation, run the code in a controlled environment (container/VM) so filesystem access is limited. Finally, clarify what 'deploy' means for your workflow — the skill saves models locally but does not include steps to publish to a remote service.Like a lobster shell, security has layers — review code before you run it.
latestvk97e74df71jdktyfd7hqj86xq58125pp
License
MIT-0
Free to use, modify, and redistribute. No attribution required.