<rss version="2.0" xmlns:dc="http://purl.org/dc/elements/1.1/" xmlns:atom="http://www.w3.org/2005/Atom"><channel><title>Hacker News: drskyle</title><link>https://news.ycombinator.com/user?id=drskyle</link><description>Hacker News RSS</description><docs>https://hnrss.org/</docs><generator>hnrss v2.1.1</generator><lastBuildDate>Thu, 19 Nov 2026 16:10:08 +0000</lastBuildDate><atom:link href="https://hnrss.org/user?id=drskyle" rel="self" type="application/rss+xml"></atom:link><item><title><![CDATA[Show HN: Hodos, a crowdsourced road-hazard map built from phone motion]]></title><description><![CDATA[
<p>Article URL: <a href="https://hodos.devilabs.dev/">https://hodos.devilabs.dev/</a></p>
<p>Comments URL: <a href="https://news.ycombinator.com/item?id=50403907">https://news.ycombinator.com/item?id=50403907</a></p>
<p>Points: 1</p>
<p># Comments: 0</p>
]]></description><pubDate>Sat, 07 Nov 2026 12:19:00 +0000</pubDate><link>https://hodos.devilabs.dev/</link><dc:creator>drskyle</dc:creator><comments>https://news.ycombinator.com/item?id=50403907</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=50403907</guid></item><item><title><![CDATA[Show HN: CloudSlash – Graph-based AWS cleaner with Terraform undo]]></title><description><![CDATA[
<p>Hi HN,<p>I’ve built an open-source CLI tool that identifies wasted cloud resources (AWS) , CLoudSlash , and, more importantly, it helps you delete them without the fear of breaking production.<p>The Problem:
Most "cloud cost" tools give you a CSV of "unused" resources. But deleting them is terrifying. Is that unattached EBS volume actually a production database backup? Is that idle NAT Gateway the only route for a silent legacy lambda?<p>I realized that detection is easy, but safe remediation is an unsolved distributed systems problem.<p>The Solution:
CloudSlash builds an in-memory Directed Acyclic Graph (DAG) of your infrastructure to understand dependencies, not just utilization metrics. It effectively treats your infrastructure as a graph traversal problem.<p>Key Technical Features:<p>1. The "Lazarus Protocol" (Infrastructure Undo Button):
   If you choose to delete a resource (e.g., a Security Group), CloudSlash doesn’t just make the API call. It first:<p>- Snapshots the live configuration.
- Generates a restore.tf file with the HCL definition.
- Generates a precise terraform import command.<p>If you make a mistake, you can resurrect the resource and bridge it back into your Terraform state in seconds. It turns "destructive" operations into reversible transactions.<p>2. Graph-Based Forensic Engine:
   Instead of simple rules ("CPU < 5%"), it builds a dependency graph.<p>- Example: It won't flag a NAT Gateway as waste—even if it has 0 bytes of traffic—if it detects an active Route Table creating a "FlowsTo" edge targeting it.
- The engine is written in Go, using custom string interning and contiguous memory layouts for the graph nodes to handle large enterprise accounts (10k+ resources) without GC thrashing.<p>3. Bin Packing, Not Just "Right-Sizing":
   For compute optimization, it uses a 2D Bin Packing algorithm (Best Fit Decreasing) to simulate how your current pods/workloads would fit into modern instance types. It visualizes the "fragmentation" of your clusters rather than just looking at average CPU usage.<p>Why Use This?<p>- Locally Run: No SaaS, no API keys leaving your machine.
- Heuristics Catalog: All detection logic is open source. You can see exactly why an RDS instance was flagged (e.g., Connections == 0 for 7 days AND SnapshotCreated == true).
- TUI: It includes a terminal UI (using Bubble Tea) for interactive graph exploration.<p>Repo: <a href="https://github.com/DrSkyle/CloudSlash" rel="nofollow">https://github.com/DrSkyle/CloudSlash</a><p>I’m eager for feedback on the graph traversal logic and the "Lazarus" restoration flow.<p>:) DrSkyle</p>
<hr>
<p>Comments URL: <a href="https://news.ycombinator.com/item?id=46603968">https://news.ycombinator.com/item?id=46603968</a></p>
<p>Points: 3</p>
<p># Comments: 0</p>
]]></description><pubDate>Tue, 13 Jan 2026 17:08:12 +0000</pubDate><link>https://news.ycombinator.com/item?id=46603968</link><dc:creator>drskyle</dc:creator><comments>https://news.ycombinator.com/item?id=46603968</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=46603968</guid></item><item><title><![CDATA[Show HN: CloudSlash – Find AWS waste and generate Terraform state rm commands]]></title><description><![CDATA[
<p>We've all been there: You find an unused NAT Gateway costing $45/mo. You delete it in the AWS console to stop the billing immediately. But the next time you run terraform plan, it fails because of state drift. Now you have to manually run terraform state rm or import it back to fix the drift. It's tedious, so often we just leave the waste running.
I built CloudSlash to automate the cleanup and the state surgery. It’s written in Go (using BubbleTea for the TUI) and solves two engineering problems:
1. Finding "hollow" resources (the graph). Most cost tools just check CloudWatch metrics (CPU < 5%). That creates too much noise. Instead, I build an in-memory graph of the infrastructure to find structural waste. Example: An "Active" ELB. It has healthy targets, so metrics look good. But if you traverse the graph (ELB -> Instance -> Subnet -> Route Table), you might see the Route Table has no path to an Internet Gateway. The ELB is functionally dead, even if AWS reports it as "healthy."
2. The state mapping. Deleting the resource in AWS is easy. The challenge is mapping a physical ID (e.g., nat-0a1b2c) back to its Terraform address (e.g., module.vpc.aws_nat_gateway.public[0]) so you can remove it from the state file programmatically. I wrote a parser that reads your local .tfstate, handles the complex JSON structure (including nested modules and for_each outputs), and generates a remediation script.
It outputs a shell script (fix_terraform.sh) that runs the necessary terraform state rm commands for you. It never writes to your .tf files directly—it just hands you the script to review and run.
The core logic, scanner, and TUI are open source (AGPLv3). I charge a one-time license for the feature that auto-generates the fix scripts for developers , but the forensic analysis/detection is free.
Repo: <a href="https://github.com/DrSkyle/CloudSlash" rel="nofollow">https://github.com/DrSkyle/CloudSlash</a></p>
<hr>
<p>Comments URL: <a href="https://news.ycombinator.com/item?id=46494848">https://news.ycombinator.com/item?id=46494848</a></p>
<p>Points: 5</p>
<p># Comments: 0</p>
]]></description><pubDate>Mon, 05 Jan 2026 02:58:35 +0000</pubDate><link>https://news.ycombinator.com/item?id=46494848</link><dc:creator>drskyle</dc:creator><comments>https://news.ycombinator.com/item?id=46494848</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=46494848</guid></item></channel></rss>