https://hellogxp.github.io/tags/llm-agents/Recent content in LLM-Agents on Xueping's BlogHugoen-usThu, 10 Sep 2026 00:00:00 +0000https://hellogxp.github.io/posts/the-context-bill/Thu, 10 Sep 2026 00:00:00 +0000https://hellogxp.github.io/posts/the-context-bill/Your agent is not slow because it has too many tools. It is paying three separate bills — a resident attention tax charged every turn, a compaction debt charged by the fix, and a measurement illusion that hides whether the capability package was worth anything. A cited accounting of all three, and a decision procedure keyed to which one you are actually paying.https://hellogxp.github.io/posts/runtime-observability-for-agent-skills/Sun, 30 Aug 2026 00:00:00 +0000https://hellogxp.github.io/posts/runtime-observability-for-agent-skills/An agent can return the right answer while its Skill runtime remains unobservable. A practical guide to evidence-graded reconstruction, versioned adapters, and the deterministic core that reliable Agent Skill infrastructure needs.https://hellogxp.github.io/posts/looping-is-not-reliability/Sun, 02 Aug 2026 00:00:00 +0000https://hellogxp.github.io/posts/looping-is-not-reliability/Why more agent revisions can discover more correct patches while delivering fewer of them—and how state-scoped evidence, read-only observation, and explicit admission boundaries turn a mutable coding trajectory into an explainable delivery chain.https://hellogxp.github.io/posts/agent-evaluation-regression-memory/Tue, 21 Jul 2026 00:00:00 +0000https://hellogxp.github.io/posts/agent-evaluation-regression-memory/Benchmarks compare agents. Production evaluation must remember failures. A research-informed practice guide to traces, calibrated graders, stochastic regression tests, coverage, and release gates for AI agents.https://hellogxp.github.io/posts/trust-stack/Sat, 27 Jun 2026 00:00:00 +0000https://hellogxp.github.io/posts/trust-stack/AI models have two safety switches — ‘knowing’ and ‘acting’ — that are structurally separated in deep layers. Models can perfectly say ’no’ while perfectly doing ‘yes.’ RLHF’s gradient is mathematically zero beyond the harm horizon. LLM judges silently soften verdicts with zero acknowledgment in their reasoning traces. Every audit tool we have checks the wrong axis. This isn’t a bug. It’s structural.https://hellogxp.github.io/posts/skill-routing/Thu, 25 Jun 2026 00:00:00 +0000https://hellogxp.github.io/posts/skill-routing/When LLM agents need to chain multiple skills, the hard part isn’t retrieval—it’s decomposition. Oracle experiments reveal a 65-percentage-point gap. SAD (Skill-Aware Decomposition) closes it by correcting granularity, not by boosting model capacity. Context window drops 99.9%.