← Back to Skills Marketplace
erongcao

PoliBERT Sentiment Analysis

by Yirong · GitHub ↗ · v1.0.0 · MIT-0
cross-platform ✓ Security Clean
71
Downloads
0
Stars
1
Active Installs
1
Versions
Install in OpenClaw
/install polibert-sentiment
Description
Political sentiment analysis using PoliBERTweet - a RoBERTa model pre-trained on 83M political tweets. Analyzes support, opposition, and stance toward politi...
Usage Guidance
This skill is coherent for political sentiment analysis but review these points before installing: 1) Installing will pull heavy ML dependencies and the PoliBERT model (~500MB) — expect large downloads and GPU/CPU resource usage. 2) requirements.txt pins versions and includes packages (pandas, scikit-learn) that the main script doesn't appear to need; consider installing only the packages you require to reduce footprint. 3) The code intends to fetch Reddit in read-only mode without credentials; PRAW usage here is uncredentialed but verify behavior in your environment (and rate limits). 4) test_sample.sh uses an absolute user path and activates a venv at that path — do not run it without adjusting the path to your environment. 5) The model comes from the HuggingFace handle in the SKILL.md; if provenance matters, verify the model owner and license on HuggingFace before use. 6) As with any political-analysis tool, results can be biased; validate outputs and consider ethical/privacy implications when analyzing user data or large social datasets.
Capability Analysis
Type: OpenClaw Skill Name: polibert-sentiment Version: 1.0.0 The polibert-sentiment skill is a legitimate tool for political sentiment analysis using the PoliBERTweet model from HuggingFace and Reddit data via the PRAW library. The code in polibert_sentiment.py and scrape_polymarket.py performs its stated functions without any evidence of data exfiltration, malicious execution, or prompt injection. While test_sample.sh contains a hardcoded local file path, this appears to be a remnant of development rather than a malicious indicator.
Capability Tags
cryptorequires-sensitive-credentials
Capability Assessment
Purpose & Capability
The skill's name/description (PoliBERT sentiment analysis, Reddit integration, Polymarket integration) aligns with the included code. However there are small mismatches: SKILL.md lists 'twitter' as a source but there is no Twitter API integration in the code; requirements.txt pins extra heavy packages (numpy, pandas, scikit-learn) that are not used by the main script, and the pinned package versions in requirements.txt (e.g., transformers==4.48.0, torch==2.6.0) are stricter than the SKILL.md prose (transformers>=4.18.0, torch>=1.10.2). These are plausibly benign but unnecessary dependencies and version pinning are disproportionate to the core task.
Instruction Scope
Runtime instructions and code focus on local text, batch files, and Reddit. The main script downloads a HuggingFace model at first run (explicitly documented). The SKILL.md claims Reddit read-only access without credentials; the code uses praw with client_id/client_secret set to None (intended for read-only usage) but does not explicitly call praw.Reddit(read_only=True). The skill does not attempt to read unrelated local system credentials or network endpoints beyond HuggingFace/Reddit. test_sample.sh references an absolute local path and activates a virtualenv in that path, which could surprise users if run without adjusting paths.
Install Mechanism
There is no automated install spec in the registry entry (instruction-only), which is low-risk. The package includes a requirements.txt listing heavy ML packages (torch, transformers, numpy, pandas, scikit-learn, praw). Installing these will pull large binaries (torch, transformers) and the model download (~500MB) will occur at first run — expected for this use case but resource-heavy. No downloads from unknown/untrusted hosts are present in install metadata; the model comes from HuggingFace (model name provided).
Credentials
The skill declares no required environment variables or credentials. The code attempts to use PRAW in an unauthenticated/read-only manner (client_id/client_secret set to None) so no API keys are required for its documented Reddit behavior. No credentials for unrelated services are requested.
Persistence & Privilege
The skill does not request persistent platform privileges (always:false) and does not modify other skills or system-wide settings. It only downloads model files to the user's environment on first run and writes no agent config.
How to Use
  1. Make sure OpenClaw is installed (local or Docker)
  2. Run the install command in chat: /install polibert-sentiment
  3. After installation, invoke the skill by name or use /polibert-sentiment
  4. Provide required inputs per the skill's parameter spec and get structured output
Version History
v1.0.0
Initial release with PoliBERTweet integration
Metadata
Slug polibert-sentiment
Version 1.0.0
License MIT-0
All-time Installs 1
Active Installs 1
Total Versions 1
Frequently Asked Questions

What is PoliBERT Sentiment Analysis?

Political sentiment analysis using PoliBERTweet - a RoBERTa model pre-trained on 83M political tweets. Analyzes support, opposition, and stance toward politi... It is an AI Agent Skill for Claude Code / OpenClaw, with 71 downloads so far.

How do I install PoliBERT Sentiment Analysis?

Run "/install polibert-sentiment" in the OpenClaw or Claude Code chat to install it in one step — no extra setup required.

Is PoliBERT Sentiment Analysis free?

Yes, PoliBERT Sentiment Analysis is completely free, licensed under MIT-0. You can download, install and use it at no cost.

Which platforms does PoliBERT Sentiment Analysis support?

PoliBERT Sentiment Analysis is cross-platform and runs anywhere OpenClaw / Claude Code is available (cross-platform).

Who created PoliBERT Sentiment Analysis?

It is built and maintained by Yirong (@erongcao); the current version is v1.0.0.

💬 Comments