Every processed story in chronological order, with the newest coverage first. Filter by tag, source, or score to drill in.
The paper identifies task decomposition — not retrieval — as the binding constraint in multi-skill agent planning, and SAD's single-iteration fix raises decomposition accuracy by over 32 percentage points, directly improving how reliably agents can assemble executable plans from large real-world skill libraries.
The single-token output bug fix restores correct multi-token generation for `ollama launch claude` and other coding agent workflows that were broken in prior builds.
VulnFeed brings EPSS-prioritized, lockfile-aware CVE scanning directly into MCP-compatible coding agents, replacing broad CVE noise with targeted, fix-ready alerts for a project's actual dependencies.
Claireon brings MCP-based AI automation directly into the Unreal Editor, allowing AI assistants to interact with a broad catalog of editor tools through a minimal, discoverable interface rather than requiring a large, manually curated tool list.
The paper demonstrates that source attribution is an independent axis of factuality verification — meaning standard source-blind metrics can pass answers that contain incorrect attributions, a gap ProvenanceGuard is designed to close in MCP-based agents.
The post illustrates the concrete gap between ChatGPT generating form field suggestions in chat and ChatGPT actually invoking remote MCP tools to create and configure a live form, showing what a working agentic write-action setup looks like in practice.
The benchmark exposes concrete, measurable gaps in LLM agents' ability to infer hidden world models through interaction, providing a rigorous testbed with classical algorithm baselines that quantifies how far current agents fall short of robust interactive discovery.
The `/import` command creates a direct migration path from Claude Code into Codex, while the new Bedrock authentication and encrypted credential storage extend Codex's reach to AWS-managed deployments.
Devloop addresses the self-review bias of single-model-family coding agents by routing implementation and review to different model families, automating the iterate-until-accepted loop so humans only intervene at the spec and PR sign-off stages.
The post demonstrates that replacing a high-token MCP workflow with a lightweight static tool can reclaim the equivalent of 7 or 8 full context windows per project, redirecting that capacity toward implementation rather than ticket management.