RepoJournal
Agentic Coding Agentic Coding
94 wires and counting

$ follow Agentic Coding

Keep up with Agentic Coding in about 3 minutes: what actually shipped — the commits, pull requests, releases, and security advisories that matter.

or

fair warning: these emails are deeply technical. diffs, version numbers, CVEs, benchmark deltas. if that's not your idea of a good read, this isn't your newsletter.

Folds into your digest — weekly by default, monthly if you prefer. Unsubscribe in one click.

$ status

wire 2026-08-25
stories 62
RSS Archive Reviews

© 2026 RepoJournal Home Showcase How it works Privacy

$ the-wire · showcase

CLINE'S SEARCH TOOL HAD A MEMORY HOLE THAT CRASHED THE AGENT

By RepoJournal · Filed · About Agentic Coding · Composed from the cited sources · methodology

Saoud Rizwan github-actions[bot] eyurtsev tofarr georgeglarson dependabot langchain-oss-model-profiles[bot]
10 people shipped this
Saoud Rizwan @saoudrizwan 2 cited github-actions[bot] @github-actions[bot] 1 cited eyurtsev @eyurtsev 1 cited tofarr @tofarr 1 cited georgeglarson @georgeglarson 1 cited dependabot @dependabot langchain-oss-model-profiles[bot] @langchain-oss-model-profiles[bot] ShashwatXD @ShashwatXD VascoSch92 @VascoSch92 Mikołaj Kondratek @mkondratek

A single-line JSON file could OOM your entire Cline agent, and the fix shipped last night in a string of hotfixes.

Cline pushed a string of four fixes that matter for anyone running the VS Code extension or CLI: first, search_codebase no longer buffers gigabytes of ripgrep stdout into one string, which killed agents on giant single-line trace dumps [1]. Before you get excited about the new default skill catalog in OpenHands , know that Cline also removed per-tool MCP auto-approve checkboxes that were no-ops, so your 'Use MCP servers' toggle is now the only gate [2]. And in a quiet move that could bite you, they disabled the todo tool and hid the Agenda UI in the desktop app while keeping schedule-only tasks , which rewrites how you plan agent runs if you rely on that view. Over in langchain, 1.3.17 shipped with a real improvement: you can now frame custom HITL rejection reasons, which means your human-in-the-loop flows can fail with intent [3], and the Anthropic gateway response metadata finally flows through for tracing [4]. OpenHands fixed the API key leak on cloud by hiding it for their provider [5], and the LLM-switching toggle is now reachable in Agent settings instead of being invisible behind a schema [6].

Action items

References

  1. [1] fix(core): prevent search_codebase from crashing the process on giant single-line files (#13525) ↗ cline/cline
  2. [2] chore(vscode): hide per-tool MCP auto-approve checkboxes behind a flag (#13522) ↗ cline/cline
  3. [3] langchain==1.3.17 ↗ langchain-ai/langchain
  4. [4] feat(anthropic): surface gateway response metadata ↗ langchain-ai/langchain
  5. [5] fix: hide API key for OpenHands provider on cloud ↗ OpenHands/OpenHands
  6. [6] feat(settings): expose the LLM-switching toggle in Agent settings ↗ OpenHands/OpenHands

$ ls ai-agents/reviews/ # the step-back reads

the week in review Claude Code RCE via malicious git executable the month in review OpenCode npm installs and Claude Desktop file bug
Full archive → Weekly reviews →

Quick answers

What shipped in Agentic Coding on August 25, 2026?
A single-line JSON file could OOM your entire Cline agent, and the fix shipped last night in a string of hotfixes. In total, 30 commits, 29 pull requests, and 3 releases landed.
Who contributed to Agentic Coding on August 25, 2026?
10 developers shipped this update, including github-actions[bot], dependabot, eyurtsev, langchain-oss-model-profiles[bot], ShashwatXD, VascoSch92, tofarr, and georgeglarson, and 2 more.
What were the notable Agentic Coding updates?
fix(core): prevent search_codebase from crashing the process on giant single-line files (#13525), chore(vscode): hide per-tool MCP auto-approve checkboxes behind a flag (#13522), and langchain==1.3.17.
♥

$ log Agentic Coding --recent

Goose 1.56 ships subagent controls; Cline fixes dropped prompts and stalled resumes 240 shipped Claude thinking text loss, AWS signing for nonempty bodies 299 shipped Codex scopes MITM credential substitution, Gemini CLI stops clobbering read-only files 276 shipped Codex stops sleeping through your input; gemini-cli 0.65.0 lands 294 shipped

$ log --all # elsewhere on the wire

Alignment & Post-training MILES ships a 25-benchmark eval panel and simpler config presets Interpretability J-space direction loss and Jacobian lens joint ablation fixed in v4.3.0

$ ls showcases/

Safety Evals AI Red Teaming Agent Security Interpretability Alignment & Post-training

$ init your-desk

RepoJournal can write one of these for any repo you care about. Sign in with GitHub, pick your repos and sources, and your desk gets a wire like this one every morning.