iCreatAI News
Back to main site
Skip to content
Trending storyWatching

AI coding agents lose context after PRs and how to address it

1 reports1 sources2 days agoUpdated

Understand this story

AI overview

sonichigo pointed out in a Hacker News post that after AI coding agents open a PR, evidence generated in delivery stages such as build failures, vulnerability scans, and latency regressions often cannot flow back into the engineering workflow, leading to context loss. The article describes using the Code Context Engine and the Software Delivery Knowledge Graph to each provide half of the context. But connecting the two is described as planned work, not a shipped feature. What is currently available is Cosmos. In local mode, it runs the Auggie CLI as an MCP server, which suits active development; in remote mode, it connects to a hosted MCP and can retrieve context across all repositories.

AI overview based on source reports · Updated 2 days ago

Latest developmentNov 12 13:35
Why AI coding agents lose context after PRs, and how to preserve it

Coverage timeline

Follow the coverage to explore different perspectives.

Nov 12
  1. Hacker News AI
    Why AI Coding Agents Lose Context After the PR, and How to Keep It

    AI coding agents can write and test code but lose the delivery evidence produced after a pull request—failed builds, security findings, latency regressions—unless it is carried back into the workflow. The post describes Harness's Code Context Engine (from its Augment Code acquisition) for code context and its Software Delivery Knowledge Graph for delivery context, citing Augment's benchmark of 80%, 71%, and 30% improvements and Harness's internal estimate of about 12K tokens per graph query versus 250K+ via raw APIs. It presents a CVE workflow combining both, notes the two-way link between them is planned rather than shipped, and covers governance via policy as code, RBAC, and approval gates.

Activity for this story

Current comparable activity
3
Peak comparable activity
10Nov 12 15:00
Change in the past 24 hours
-51%

Trends compare the same fully observed participants, which may cover fewer participants than current activity. Move the pointer or tap the chart to inspect hourly activity; use the left and right arrow keys with a keyboard.