SShortSingh.
TopTechnologyAI & MLStartupsProgrammingScienceBusinessFinanceCrypto & Web3World
Back to feed
Programming

Proven: a coding agent's ticket closes only after a Kane test fails, then passes

0
·16 views

A developer built Proven, a ticket board for the Kane CLI hackathon, to stop coding agents from marking work done without evidence. Each ticket's plain-English Kane test is frozen and must fail on the current app before any change is made. After the agent's fix, Kane must pass the test, and a receipt linking the red and green runs is recorded. Closed tickets are rechecked with fresh runs, which caught a regression in the demo shop where a second ticket broke an earlier one. The tool currently supports only static web apps, and receipts are tamper-evident rather than tamper-proof.

Read the full story at DEV Community

This is an AI-generated summary. ShortSingh links to the original source for the complete article.

Discussion (0)

Log in to join the discussion and vote.

Log in

Related stories

0
ProgrammingHacker News ·

Lovable-built apps reportedly exposed passport details and over 200,000 private messages

A security research report published on crissyfield.de says apps built with the Lovable platform leaked sensitive user data. The reported exposures include passport details and more than 200,000 private messages. The findings were shared on Hacker News, where the post drew modest attention. The source material provided is limited to the link and metadata, so the full technical details, affected scope, and company response are not confirmed here. Readers should consult the original research for specifics.

0 comments Read more at Hacker News
0
ProgrammingDEV Community ·

Source-traceable retrieval design for versioned insurance claims intake records

The article argues that insurance claims intake should use a staged retrieval architecture with explicit collections, bounded queries, and source context that pinpoints the exact document revision. It recommends a deterministic revision gate before vector search, so only eligible revisions are ranked and superseded evidence cannot surface. Each accepted document revision is treated as an immutable retrieval unit, with a separate pointer marking the current version and deliberate deletion for withdrawn records. The author stresses that these correctness rules belong to the application rather than any specific vector database provider. A worked example involving a corrected police report illustrates how stale or partially indexed revisions could lead to wrong answers.

0 comments Read more at DEV Community
0
ProgrammingDEV Community ·

Queueing tasks replaces rejections, but the system still pushes work to workers

The author argues that a full worker pool only reports that no capacity is available because nothing records the work still waiting. Placing tasks into a queue and reporting each user's position turns rejections into waiting, which is a better experience at peak load. However, the scheduler still checks each worker for availability and then assigns tasks, which keeps the race condition and the cost of polling. The author notes that true pull-based scheduling, where free workers take tasks from the queue themselves, is needed to remove those problems. The post also points out that a VIP flag is passed along but never used by the scheduler.

0 comments Read more at DEV Community
0
ProgrammingDEV Community ·

WiFi presence sensors can misread rooms after furniture moves, so recalibration matters

A WiFi presence sensor's main failure mode is often moved furniture rather than people, since it relies on a baseline of the room's radio field. Changes like a relocated bookshelf, metal cabinet, or new router alter that baseline and can cause false continuous-activity readings. The developer's approach treats large, persistent shifts as recalibration events and shows a 'no signal' state when changes are too unclear to classify. The device cannot distinguish pets or fans from people and does not coordinate across multiple units, and it is not a medical device. Its public demonstration board reports live data from an occupied flat.

0 comments Read more at DEV Community