signread: an open-weight model that reads your wallet signature before you sign it

Description and use

signread uses a local Qwen 2.5 1.5B model to analyze a crypto wallet signature, providing a risk verdict and plain‑language explanation so users can avoid malicious transactions.

Problem

Crypto wallet users often sign transactions they cannot read, leading to drains such as free listings, blanket approvals, or hidden fees.

Outcome

signread provides a deterministic verdict and a plain‑language explanation for wallet signatures, enabling users to avoid signing harmful transactions without relying on external APIs.

Workflow

• Paste the transaction payload into the GitHub Actions button or CLI
• The runner installs Ollama and pulls the Qwen 2.5 1.5B model
• The payload is decoded deterministically to extract structured findings
• The model is prompted with the findings and generates an explanation in the user’s language
• The verdict is displayed in the run summary
• If the verdict is HIGH, the CI fails and the transaction is blocked

Evidence

signread takes that payload and turns it into one verdict (HIGH, MEDIUM or LOW), the facts behind it, and a short explanation in plain language from a local open-weight model.

Source

The runner installs Ollama, pulls Qwen 2.5 1.5B (Apache-2.0, open weights) and runs it on CPU. No API key, no account, and the payload never leaves the runner.

Source

A deterministic decoder, covered by tests, computes the verdict. The model only gets the structured findings, is told to treat quoted text as untrusted data, and is not allowed to change the verdict.

Source