CtrlK
BlogDocsLog inGet started
Tessl Logo

tessl-labs/pr-review-guardrails

Evidence-first pull request review with independent critique, selective challenger review, and human handoff.

89

1.36x
Quality

92%

Does it follow best practices?

Impact

89%

1.36x

Average score across 43 eval scenarios

SecuritybySnyk

Risky

Do not use without reviewing

Overview
Quality
Evals
Security
Files

risk-routing.mdrules/

Risk Routing

The risk lane determines which skills run, how findings are surfaced, and how much time the tile spends.

Lane routing

  • Green: lightest viable workflow (evidence + 1 review pass + synthesis)
  • Yellow: standard workflow (evidence + review + optional challenger + synthesis + handoff)
  • Red: full workflow with mandatory human review

Challenger triggering

Use challenger review selectively based on risk lane or reviewer uncertainty. Never skip the evidence-build stage regardless of risk lane.

Confidence rounding

When the risk classifier's confidence is low, route to the next higher lane. False red is cheap. False green is dangerous.

Performance budgets

  • Green: synchronous, under 2 minutes
  • Yellow: synchronous, under 5 minutes
  • Red: may run async, under 10 minutes, notifies when ready

If the tile is slower than these budgets, developers will skip it.

README.md

tile.json