Persist verify outcomes to SQLite

[ionfwsrijan]

Description

The verify step's pass/fail result is a free training label — it tells us whether a proposed fix actually worked. This is the most valuable signal in the entire ML pipeline. This issue persists it.

What to implement

Add to backend/app/db.py:

CREATE TABLE IF NOT EXISTS verify_outcomes (
    id                      TEXT PRIMARY KEY,
    job_id                  TEXT NOT NULL,
    passed                  INTEGER NOT NULL,  -- 1 = passed, 0 = failed
    new_issues_introduced   INTEGER DEFAULT 0,
    verified_at             TEXT DEFAULT (datetime('now'))
);

After /verify completes, write a row to verify_outcomes with:

• job_id from the request
• passed: 1 if no new critical/high findings were introduced, 0 otherwise
• new_issues_introduced: count of net-new findings vs pre-fix baseline

Acceptance criteria

• verify_outcomes table is created on startup
• Every /verify call writes an outcome row
• passed is always 0 or 1, never NULL

---

[sasidaran-99]

Hi, I'd like to work on this issue. I'll implement the SQLite table for verification outcomes and persist the /verify results as described.

[ionfwsrijan]

Hey @sasidaran-99!

You've been assigned to this issue.

Before you start, a couple of quick things:

⭐ Star the repo — it helps the project grow and supports the roadmap
🔀 Fork it — work on your fork and open a PR from there when ready

Feel free to drop a comment here if you run into anything or have questions while implementing. Happy building!

[sasidaran-99]

Hey @sasidaran-99!

You've been assigned to this issue.

Before you start, a couple of quick things:

⭐ Star the repo — it helps the project grow and supports the roadmap
🔀 Fork it — work on your fork and open a PR from there when ready

Feel free to drop a comment here if you run into anything or have questions while implementing. Happy building!

Is there any specific deadline by which I need to submit the PR?

[ionfwsrijan]

@sasidaran-99 Yes. You got 2 days and its a beginner friendly issue. Do remember to star the repo

[sasidaran-99]

@sasidaran-99 Yes. You got 2 days and its a beginner friendly issue. Do remember to fork the repo

Got it

[ionfwsrijan]

@sasidaran-99 Star the repo. It helps the project to grow

[sasidaran-99]

@ionfwsrijan ,I started investigating this issue on the current main branch.

I couldn't find backend/app/db.py or any existing SQLite/database layer. The /verify endpoint currently lives in backend/app/main.py, calls sandbox/verify.py, and returns a VerifyResponse containing only ok and checks.

Could you confirm whether this issue targets a different branch/version, or point me to where the database code is located?

[sasidaran-99]

@ionfwsrijan
I also wasn't able to find any existing SQLite setup or database initialization code.

Should I create the database layer and db.py module from scratch as part of this issue, or is there another branch/version where these components already exist?