2026-05-10

• #0010

  #0010Infra & SaaSGitHub Trending Weeklylast week

  `InsForge`, open-source backend stack built for coding agents

  60radar

  InsForgeBackend stack — all-in-one backend layer for agents

  Puts auth, Postgres, storage, functions, hosting, and an OpenAI-compatible gateway behind one agent-friendly control layer. Worth tracking if you want a self-hostable alternative to stitching together Supabase + Vercel + AI proxy logic.

  • —Agents can operate it through an MCP Server or CLI + Skills, so backend setup and debugging become callable tools instead of manual dashboard work.
  • —The product surface is broad: Authentication, Database, Storage, Model Gateway, Edge Functions, site deployment, plus private preview compute.
  • —The Model Gateway exposes an OpenAI-compatible API across multiple providers, which lowers switching cost for LLM-backed apps.
  • —Self-hosting starts with Docker Compose and Node.js, making local end-to-end backend testing possible before committing to a hosted stack.

  Source: github.com/InsForge/InsForgeRead original →
• #0009

  #0009Agents & toolsGitHub Trending Weeklylast week

  `Ouroboros`: spec-first Agent OS for replayable AI coding

  60radar

  OuroborosAgent OS — spec-first AI coding workflow

  Adds a local-first runtime that turns agent coding into a replayable, policy-bound workflow. Worth testing if you want less prompt thrash and more verifiable runs across multiple coding CLIs.

  • —Uses a fixed loop of interview -> crystallize -> execute -> evaluate -> evolve, pushing requirements capture ahead of code generation.
  • —Targets multiple runtimes including Codex CLI, Claude Code, OpenCode, Hermes, Copilot CLI, and Kiro, reducing lock-in.
  • —The pitch is a 3-stage automated evaluation gate plus an immutable seed spec, so architecture drift and manual QA get constrained early.
  • —Setup is oriented around local tooling and MCP registration, with ooo interview "..." as the entry point instead of freeform prompting.

  Source: github.com/Q00/ouroborosRead original →
• #0008

  #0008Agents & toolsopencode_releaseslast week

  `OpenCode` `v1.14.46` hardens config, API schema, and Plan Mode

  50radar

  OpenCodeOpen-source coding agent — local config and MCP extensibility

  Startup and SDK reliability improved across config edits, query typing, and old session data. Small release, but worth taking if you rely on HTTP API, MCP, or subagents.

  • —A built-in customize-opencode skill reduces breakage when editing config, lowering startup risk.
  • —OpenAPI and SDK generation now handle numeric and boolean query params correctly for session, file, and workspace-routed endpoints.
  • —Legacy numeric values and negative token counts in old sessions no longer break message loading or Desktop startup.
  • —MCP tool discovery now survives broken outputSchema refs from servers, which cuts failure from third-party integrations.
  • —A Plan Mode security bypass was closed so subagents can no longer ignore parent deny rules.

  Source: github.com/anomalyco/opencode/releases/tag/v1.14.46Read original →
• #0007

  #0007Agents & toolsopencode_releaseslast week

  `OpenCode` `v1.14.45` fixes path rules, session loading, and SDK errors

  50radar

  OpenCodeOpen-source coding agent — ships CLI, TUI, and desktop

  This patch cleans up permission matching, session recovery, and API error handling. Low drama release, but worth taking if you use worktrees, custom providers, or the desktop app.

  • —Read allowlists and denylists now match worktree-relative paths correctly, fixing permission drift in multi-worktree setups.
  • —Workspace-routed HTTP endpoints accept valid directory and workspace query params again, so automation against the local API is less brittle.
  • —The TUI now reports all failed bootstrap requests and avoids a crash when message loading fails, which shortens debugging time.
  • —Desktop restores older migrated sessions even when diff metadata or patches are missing, reducing breakage after upgrades.
  • —throwOnError: true now throws a real Error with the server message and keeps the response body in cause, making SDK failures easier to inspect.

  Source: github.com/anomalyco/opencode/releases/tag/v1.14.45Read original →
• #0006

  #0006Agents & toolsr/ClaudeAIlast week

  `Autoharness` lets `Claude Code` tune its own agent harness

  50radar

  AutoharnessHarness tuner — validates and keeps only score gains

  A layer above agent setup is emerging: tools that rewrite prompts, scoring, and runtime context, then keep only eval-proven wins. Concrete benchmark lifts make this more than workflow bragging, and the repo is worth skimming if you run custom agents.

  • —On tau2-airline, it reports +40.7% from best-of-N skillbook scoring with an LLM judge; evaluator design is moving into the product surface.
  • —Another +24.1% came from reflector hyperparameter changes like temperature and max subagent calls, so simple harness knobs still have large headroom.
  • —Injecting runtime context on every step added +22.2%; step budget, recent tool calls, and recent results materially changed outcomes.
  • —Flow is minimal: install, point Claude Code at GUIDE.md, propose harness edits, run evals, and keep only score-improving changes.
  • —Because gains are benchmark-specific, the useful takeaway is the loop: mutate harness, eval immediately, and ship only measured improvements.

  Source: www.reddit.com/r/ClaudeAI/comments/1t8cn9y/claude_improvRead original →
• #0005

  #0005Agents & toolsr/ClaudeAIlast week

  One Escaping Bug Turned `rd /S /Q \` Into a 717 GB Windows Wipe

  40radar

  A cross-shell quoting failure turned a folder delete into C:\ root deletion in under two minutes. Useful reminder: never pipe destructive ops through multiple parsers when a native command plus -WhatIf exists.

  • —The intended target was C:\Users\ADMIN\Desktop\WIP, but cmd /c plus zsh/tmux/PowerShell collapsing changed it to rd /S /Q \.
  • —Damage reached about 717 GB before interruption; only a separate HDD backup prevented real loss, which makes backup isolation the real safety net.
  • —Using PowerShell Remove-Item -Path 'C:\absolute\path' -Recurse -Force -WhatIf would have exposed the broken path before execution.
  • —The practical lesson is operational, not model-specific: echo the fully expanded command and avoid routing deletes through cmd when PowerShell can run them directly.

  Source: www.reddit.com/r/ClaudeAI/comments/1t923er/i_deleted_a_gRead original →
• #0004

  #0004Infra & SaaSGeekNewslast week

  `AWS` Virginia Outage Resolved After `US-East-1` Overheating Incident

  40radar

  US-East-1AWS region — its primary Northern Virginia region

  A cooling-related failure in `US-East-1` hit major customer-facing services, not just internal AWS dashboards. If anything still depends on a single region, this is another reminder that multi-region failover is cheaper than surprise downtime.

  • —The trigger was data center overheating in US-East-1, turning a routine status alert into a real regional availability event.
  • —Named impact reached services like Coinbase and FanDuel, which shows the blast radius reached end-user transactions, not background jobs only.
  • —AWS marked recovery at 3:29 PM ET Friday. Even a resolved incident matters if your stack, auth, or webhooks are pinned to one region.

  Source: news.hada.io/topic?id=29337Read original →
• #0003

  #0003Agents & toolsopencode_releaseslast week

  `Opencode` `v1.14.44`, fixes upgrade failures in existing workspaces

  40radar

  OpencodeOpen-source coding tool — workspace-based dev workflow

  A patch fixes existing workspace upgrades breaking when the time_used field is introduced. Relevant if you update often; otherwise urgency is low.

  • —The fix targets upgrade failures in existing workspaces, not new installs, so the blast radius is narrow.
  • —The trigger was adding the time_used field, which means schema or state migrations were the failure point.
  • —If your team pins Opencode, this is a safe maintenance update rather than a feature-driven release.

  Source: github.com/anomalyco/opencode/releases/tag/v1.14.44Read original →
• #0002

  #0002Agents & toolsopencode_releaseslast week

  `OpenCode` `v1.14.43`, fixes `ACP` image attachments and auth-loader API compatibility

  40radar

  OpenCodeOpen-source coding agent — supports session replays

  Restores ACP image attachments in updates and session replays, and keeps provider/config APIs working when auth loaders inject non-JSON options. A small release, but worth taking if you rely on multimodal traces or custom auth.

  • —ACP updates and session replays now include tool image attachments. Screenshot-driven agent runs become easier to audit and debug.
  • —Provider and config API responses no longer break when auth loaders add non-JSON provider options. Custom auth integrations get safer.
  • —This is a Core bugfix-only release with one contributor credited. Low migration risk, mostly operational cleanup.

  Source: github.com/anomalyco/opencode/releases/tag/v1.14.43Read original →
• #0001

  #0001OtherGeekNewslast week

  `NixOS` Secrets: Plaintext in `Nix store` Is Readable

  40radar

  NixOSLinux distro — declarative, reproducible system config

  Putting secrets in Nix config, private Git, or git-crypt plaintext still leaks them through Nix store. sops-nix is the practical baseline if you deploy with NixOS; otherwise your server access model is already broken.

  • —Secrets embedded in Nix config can end up in Nix store, which is world-readable on the machine; repo privacy does not help after deploy.
  • —Keeping values in a private repo or decrypting with git-crypt before build still leaves plaintext artifacts in the store, so the leak point moves to runtime.
  • —sops-nix uses .sops.yaml rules and encrypted-file editing flow, giving a cleaner path than ad hoc secret injection for NixOS setups.

  Source: news.hada.io/topic?id=29328Read original →