Bugbot PR Proof Demo

Verify AI-generated PRs with proof, not trust.

Bugbot does not just review AI-generated code. It proves whether the outcome can be trusted.

Every pull request analysis produces a proof chain: read-only proof, a learning archive, historical insight, a final outcome, and a contract-validated result. This page shows a sample of that chain using sanitized demo data.

Static sample data only — no live GitHub execution from this page.

This is a static proof demo. All data shown is sanitized sample output. No live PR analysis is performed from this page.

The proof chain

PR → Proof → Learning → Insight → Outcome → Contract

Each step produces a machine-readable record. The final outcome is validated against bugbot.module.contract.v1 before any downstream consumer can trust it.

Sample outcome

ACCEPTED Outcome verified — PR is safe to proceed

PR example-org/example-repo #42

Runner bugbot-outcome-demo

Status ACCEPTED

Decision ACCEPTED

Trusted true

Summary No contamination detected. All safety checks passed.

Action Safe to proceed with human review.

Proof chain

Every step is verified

01 Proof Read-only GitHub analysis — writesDetected: false · 4 files changed PASS
02 Learning Archive written · 1 record stored · 5 historical records read PASS
03 Insight Dominant decision: APPROVED · Dominant risk: INFO · No weak spots detected PASS
04 Outcome Status: ACCEPTED · trusted: true · Proceed with human review PASS
05 Contract bugbot.module.contract.v1 · valid: true · errors: 0 PASS

Machine-readable output

Sample JSON — the full payload

The outcome and contract validation are both machine-readable JSON. Downstream consumers read these directly — no parsing of human text required.

Contract validation

bugbot.module.contract.v1

Every Bugbot outcome is validated against a machine-enforceable contract before any downstream system can trust it.

Contract

bugbot.module.contract.v1

Version

1.0

Status

ACCEPTED

Valid

true

Errors

Privacy safe

true

GitHub write safe

true

GENESIS core safe

true

Privacy & safety

Verifiable guarantees, not promises

These flags are enforced by the contract validator — not asserted manually. Any violation causes the outcome to be rejected as INCONCLUSIVE.

Privacy

Raw source code: never stored
Raw diff content: never stored
File contents: never stored
Sensitive values: never stored
Stores only metadata: YES

GitHub safety

GitHub read-only: YES
No GitHub writes performed: YES
No merge performed: YES
No PR created by Bugbot: YES

What this demo is not

This page shows a static proof demo only. It does not execute live analysis.

No live GitHub execution from this website
No automatic merge or PR comments
No status checks posted to GitHub
No live PR data — all data shown is fictional sample output
Bugbot is a prototype proof layer, not a production security engine
Running against a real PR requires a local Bugbot verifier environment

← Back to AI Code Safety Gate demo

Questions about the proof model?

Email hello@aicodesafety.com