Who changed it?
Connect code changes to the human owner, AI agent and delivery system involved.
Public trust surface
Gilded Code is building devOS: a governance layer for teams adopting AI coding agents. It turns pull request activity into reviewable evidence packs that show who approved the work, which controls ran, and what changed.
The audit questions
Connect code changes to the human owner, AI agent and delivery system involved.
Show the controls, permissions and review gates that were active for the work.
Produce signed artifacts that can be inspected after the pull request has moved on.
Produce evidence that security, compliance and platform teams can consume.
devOS
devOS focuses on the evidence around delivery, not on replacing source control, CI or identity platforms. It is designed to sit beside GitHub, CI systems and enterprise controls as the governance record for software change.
Collects PR metadata, review decisions, checks and agent attribution.
Maps the work against delivery controls and flags missing evidence.
Stores a compact evidence pack that survives branch cleanup and tool churn.
Gives platform, security and compliance teams a shared record to inspect.
Trust posture
Gilded Code demonstrates a practical evidence workflow for governed AI-assisted delivery.
The product narrative is grounded in identity, permission boundaries, review gates and durable proof of agent-assisted work.
The workflow is intentionally GitHub-native while keeping the evidence contract usable by identity, ticketing and compliance systems.
Evidence pack
The PR evidence workflow gives customers a clear record of review decisions, policy checks and supporting artifacts for AI-assisted software delivery.
| Field | Purpose |
|---|---|
| Attribution | Connects human and agent activity to the PR. |
| Controls | Shows which governance checks ran and whether they passed. |
| Artifacts | Records evidence, hashes and generated review outputs. |
| Verdict | Summarizes whether the PR is ready for review, escalation or remediation. |
Design partners
We are looking for teams adopting AI coding agents who need a defensible way to show review, attribution and governance evidence around software change.