Rescanning Files

TIP

Consider running a rescan after dismissing findings to improve the quality of future results.

Rescanning allows you to reanalyze your codebase with updated changes and configurations.

How to Rescan

Recanning is the same process as the initial scan:

1. Go to your Project page
2. Once the scan configuration is set click Scan Project

Why Rescan?

• New code changes – run a scan after you merge PRs, refactor, or add new contracts.
• Knowledge-base updates – after uploading docs, rescan so the LLM uses the new context.
• Rule tweaks – run a new scan after adding or editing rules.
• Scan-mode switch – toggle between DEFAULT and LOW_NOISE.
• Threshold tuning – experiment with different severity / count gates.
• Verify fixes – confirm previously flagged findings are resolved.
• Model improvements – benefit from the newest AI engine without waiting.

Do findings change between runs?

Yes, Almanax uses a machine learning based method and LLM's are nondeterministic, therefore we cannot guarantee consistency between runs. However, you can refine scan results based on Dismissals. For instance, if you dismiss a particular finding as a False Positive, Nit, or Won't Fix type, subsequent scans of the same file will omit that dismissed finding, ensuring cleaner, more relevant output.

Example Rescan Workflow

1. Initial Scan: The scanner flags a hard-coded AWS secret key inside src/config.js.
2. User Action: You rotate the key, move secrets to environment variables, and dismiss the original finding as Fixed.
3. Rescan: The updated codebase is re-analyzed and the secret-key finding no longer appears.

This approach helps reduce noise and improves overall output relevance as findings are addressed.