Claude Tool Calls, Codex SSD Wear, and Python-Shell Agents
Tuesday, August 11, 2026 · 10 min
Claude Sonnet 4.5 on Amazon Bedrock now preserves trailing newlines in tool-call parameters, while Codex’s SQLite logs exposed SSD-endurance risk and Prime Agent’s Python-shell design points at a different answer to context rot.
Listen
Show notes
Claude Sonnet 4.5 on Amazon Bedrock now preserves trailing newlines in tool-call parameters, while Codex’s SQLite logs exposed SSD-endurance risk and Prime Agent’s Python-shell design points at a different answer to context rot.
In this episode
- Anthropic Claude tool use - Amazon Bedrock — Amazon
# Anthropic Claude tool use **Warning** Several functions below are offered in beta as indicated. These features are made available to you as a "Beta Service" as defined in the AWS Service Terms. It is subject to your Agreement with AWS and the AWS Service Terms, and the applicable model EULA. With Anthropic Claude models, you can specify a tool that the model can use to answer a message.…
- code-yeongyu/oh-my-openagent v5.0.0-beta.4 on GitHub — Newreleases
code-yeongyu/oh-my-openagent v5.0.0-beta.4 on GitHub v5.0.0-beta.4 on GitHub latest release: v5.0.0-beta.5 19 hours ago - ea2b387 - Merge pull request #6693 from code-yeongyu/release/v5.0.0-beta.4-source-state - f065c0f - Merge pull request #6688 from code-yeongyu/fix/memory-surface-direct - 8357cb5 - docs(agents-md): refresh knowledge base against HEAD
- Codex SQLite feedback logs can write ~640 TB/year and rapidly consume SSD endurance — GitHub
# Codex SQLite feedback logs can write ~640 TB/year and rapidly consume SSD endurance Published: 2026-08-10T05:23:53+00:00 Source: github.com (github.com) Language: en ## Story # Codex SQLite feedback logs can write ~640 TB/year and rapidly consume SSD endurance - State: closed - Author: 1996fanrui - Created: 2026-06-14T21:34:35Z - Updated: 2026-08-03T17:04:32Z - Repository: openai/codex -…
“Someone posted a temporary workaround for this on X[1]. sqlite3 ~/.codex/logs_2.sqlite "CREATE TRIGGER IF NOT EXISTS block_log_inserts BEFORE INSERT ON logs BEGIN SELECT RAISE(IGNORE); END;" Also, I found that running VACUUM FULL on the sqlite file on my laptop shrunk it from…” — Hacker News (510 pts thread)
Our take: That trigger is the exact kind of engineer-brained tourniquet we respect: ugly, local, and immediately testable. But if you use it, remember you’re also choosing to throw away those inserts; VACUUM gets your space back, it doesn’t prove the logging path is safe again.
“Shocking. Been open a week and AFAICT just silence from OpenAI. I just find it baffling. You'd think that these vendors would be very sensitive to this sort of issue. I mean, surely they have multiple agents hooked up to github monitoring potential issues and proposing fixes,…” — Hacker News (510 pts thread)
Our take: The jab lands because an AI coding company really should be allergic to unattended write amplification. The wrinkle is that the issue now lists merged fixes, so the enduring question is less “can nobody fix it?” and more “why did users have to become the SSD smoke alarm?”
“Well, everyone's bashing on OpenAI as well they should, but just a reminder, unlike Claude Code, Codex is officially available to customize here: It's fairly easy to patch.” — Hacker News (510 pts thread)
Our take: Fair counterweight: Codex being patchable changes the blast radius for teams willing to own their tooling. It still doesn’t make a default feedback log that can chew through drive endurance feel like an acceptable production posture.
- The top repo on GitHub today runs its agent inside a Python shell - DEV Community — DEV Community
The top repo on GitHub today runs its agent inside a Python shell - DEV Community Prime Intellect's Prime Agent hit number one on GitHub Trending with 2,319 stars in a single day, and its design is the notable part: instead of stuffing a task into an ever-growing conversation, it hands the model a live Python session and lets it manage its own context as code. The repository now carries 11,400…
- AI Engineering Evals: Regression Gates for Prompts, Tools, RAG [2026] - DEV Community — DEV Community
AI Engineering Evals: Regression Gates for Prompts, Tools, RAG [2026] - DEV Community Posted on Aug 10 • Originally published at kunalganglani.com # AI Engineering Evals: Regression Gates for Prompts, Tools, RAG [2026] > Originally published at kunalganglani.com — read it there for inline code, hero image, and live links. AI engineering evals testing regression gates is the unsexy…