ClawHub

Data Anomaly Detector

v2.1.0

Detect anomalies and outliers in construction data: unusual costs, schedule variances, productivity spikes. Statistical and ML-based detection methods.

⭐ 1· 3.1k·28 current·29 all-time
by@datadrivenconstruction
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
ℹ
Purpose & Capability
The name/description match the instructions and embedded Python code (statistical and ML methods for construction data). However, the SKILL.md uses pandas, numpy, scipy, etc., while the registry only declares python3 as a required binary — the skill omits required Python package dependencies, which is an inconsistency.
✓
Instruction Scope
Runtime instructions focus on processing user-provided construction data (CSV/Excel/JSON) and presenting results. The SKILL.md does not tell the agent to read unrelated system files, call unknown external endpoints, or exfiltrate data — it instructs use of user-supplied inputs.
ℹ
Install Mechanism
This is an instruction-only skill with no install spec (lowest install risk). But because the embedded code requires Python packages, the lack of an install step or dependency manifest (requirements.txt, pipenv, or similar) is a packaging gap: an agent or user may attempt to pip-install dependencies at runtime, or the skill may fail.
✓
Credentials
The skill requests no environment variables, no external credentials, and no config paths. That is proportionate to the stated purpose of local data analysis.
ℹ
Persistence & Privilege
claw.json indicates the skill requests filesystem permission (reasonable for reading user files and exporting reports). always:false and normal autonomous invocation are used. Users should be aware filesystem access allows reading paths the user provides — verify that you only provide datasets the skill should access and prefer sandboxed execution if needed.
What to consider before installing
This skill appears to implement local anomaly detection logic and does not request credentials, but there are a few practical concerns you should consider before installing: - Missing Python dependencies: The included code uses pandas, numpy, scipy, and possibly other Python packages, but the skill only declares python3. Ask the author for a requirements.txt or installation instructions, or prepare to install those packages in a controlled environment. - Filesystem access: The skill declares filesystem permission so it can read user-supplied CSV/Excel/JSON files and write exports. Only supply files you intend to analyze and run the skill in an environment you trust or sandboxed container if data is sensitive. - No install or package provenance: Because this is instruction-only, there is no packaged code to inspect beyond SKILL.md. If you will run any pip installs suggested by the agent, prefer official PyPI packages and review what will be installed. - Verify the homepage/author if you need stronger assurance: the registry metadata lists a homepage (datadrivenconstruction.io); consider confirming documentation or a source repository for reproducible installs. If you require higher assurance, ask the publisher for a dependency manifest and an installation script that targets a virtual environment or container, and confirm that the skill does not attempt to send data to external services.

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

latestvk976vzt7xq6cf0nj70rrmznn5s8161qv

License

MIT-0
Free to use, modify, and redistribute. No attribution required.
Termshttps://spdx.org/licenses/MIT-0.html

Runtime requirements

✔️ Clawdis
OSmacOS · Linux · Windows
Binspython3

Comments