Regression Provenance

Type: Blame-tracking pattern
Referenced from: openclaw-autoreview-skill

Definition

A structured approach to tracing code issues back through the full chain of responsibility: who wrote the code, who PR'd it, who merged it, and when. When automation merged the PR, the pattern identifies the human who triggered the automerge.

Roles Tracked

Role What It Means
Blamed code author Who wrote the problematic code
Blamed PR author Who opened the PR containing the code
PR merger/committer Who merged the PR (human or bot)
Current PR author Who is currently working on the code
PR/date Metadata for temporal context

When No PR Exists

If no blamed PR is traceable, the blamed commit becomes the provenance:
- Commit SHA
- Commit date
- Author username

Don't guess a merger. Don't frame missing PR metadata as a separate finding.

Automerge Identification

When the PR was merged by automation (clawsweeper[bot] or similar):

  1. Check timeline/comments for maintainer commands
  2. Look for @clawsweeper automerge, /landpr, or labels/status comments
  3. Report: automerge triggered by @login
  4. If not found: Report trigger unknown

This matters because automerge obscures the human decision-maker. Without identifying the trigger, regression analysis can't attribute responsibility.

Why This Matters

  • Accountability — Know who decided to ship the code that caused the regression
  • Pattern detection — If the same author/merger keeps introducing regressions, the pattern is visible
  • Automation transparency — Bots merge code, but humans trigger bots. The human is the responsible party.
  • Incident response — When a regression hits production, provenance tells you who to talk to

Application to Factory Pipelines

In a factory pipeline, regression provenance is a metadata enrichment step during code review:

Code review → [Provenance enrichment] → Findings with full blame chain

Every finding includes the blame chain, so the fix author knows exactly who to consult and what context to consider.