Newsletter archive
Applied Signals Weekly: 2026-08-23
Aug 17, 2026 to Aug 23, 2026
In this week’s edition
This week’s edition surfaces where agentic systems face the sharpest pressure points: reliability under coordination, evaluation discipline, and the practical mechanics that keep multi-agent behavior stable over time.
- AI4AI-Bench: Benchmarking LLM Agents in Algorithmic Design for Recursive… — Recursive self-improvement (RSI) asks whether an AI system can improve the process that produces AI systems, so that the next system inherits the improvement. That process is the training algorithm: a better objective or…
- DECOWAM: Decoupled Whole-Body World-Action Model for Legged Mobile Manipulation — Mobile manipulation requires a robot to predict how locomotion and arm motion jointly alter future observations and control. Existing world-action models, developed largely for fixed-base platforms, do not explicitly…
- MaliciousSkillBench: A Comprehensive Benchmark for Malicious Agent Skill… — Agent Skills extend LLM agents with reusable instruction packages that may also include scripts, resources, and service configuration. This creates a direct distribution channel for malicious behavior, yet existing…
- MemTrapBench: Benchmarking Cognitive Traps in LLM Memory Use — Memory has become a key component of large language models, enabling them to retain information and learn from long-term interactions. However, existing memory benchmarks mainly evaluate whether information is correctly…
- What You Can't See Is What You Learn: Restricted Evidence Visibility Favors… — Multi-module systems often expose every module to the full input. We test whether restricting evidence visibility changes which solutions gradient-based training discovers. Four-cell societies share one frozen pretrained…
- How Much Memory Does Your Agent Actually Need?
- Is your coding agent uploading all your code? — After Grok Build was caught uploading entire Git repos, we read the privacy docs for Claude Code, Codex, Cursor, GitHub Copilot, and Grok Build to compare what code leaves your machine, what gets retained, and what…
Worth watching
- evaluation + tool-use: hype or repeatable pattern?
- evaluation + planning: hype or repeatable pattern?
- evaluation + frameworks: hype or repeatable pattern?
Read the full analysis
- AI4AI-Bench and Outcome Monitors quantify where agent reliability breaks (weekly trend)
- Aug 23, 2026 MaliciousSkillBench source-disjoint eval exposes recall–false-positive tradeoffs
- Aug 22, 2026 Restricted Evidence, AI4AI-Bench and PolicyGuide
- Aug 21, 2026 MemTrapBench memory failures, ToolMaze Outcome Monitors, ECP evaluation protocol
- Aug 20, 2026 SPADE self-play environments, BayesBeliefAgent replanning, CentaurBench assistance evals
- Aug 19, 2026 Aegis runtime governance, GxP-Agent DAGs, StartupBench results
- Aug 18, 2026 BRA-Audit scheduling, Hallucination Snowball handoff gates, HarnessEval-W evidence trees
- Aug 17, 2026 Language-server retrieval, InflationAgent routing, RubricForge evaluation findings
Subscribe
Get Applied Signals weekly newsletter in your inbox