Fetching from the wire…
01
02
03
04
05
06
07
08
Public story · 2026-03-20 · source-backed
ArXiv paper 2603.14805 argues the primary bottleneck in scaling agentic development is knowledge architecture — "skills" (composable, governance-aware units encoding institutional knowledge) are the right primitive for agents, not raw docs or in-context retrieval. Directly relevant to the SKILL.md and AGENTS.md standardization trend. The framework positions skills as the mechanism for converting tacit engineering expertise into executable agent behavior.
Each link below shares sources, entities, or timing with this story.
This one annoyed me, because I've been running the losing pattern. SWE-QA (arXiv 2608.01507) compares the sub-agent grep pattern that Claude Code, Codex and Antigravity all ship by default against a pre-built semantic index over the same repository. Semantic search answered 65...
ArXiv 2603.17368 proposes evaluating safety policy before chain-of-thought generation rather than after. Models that reason first and apply safety second can be manipulated through the reasoning trace itself. Reordering substantially improves alignment without degrading benchm...
The serving system targets the practical inference-cost wall you hit when agents produce very long outputs and dense attention becomes the bottleneck (arXiv). As agent runs get longer, this is the kind of infra that decides whether serving them at scale is affordable. Directly...
arXiv 2603.15566 proposes treating commit history as machine-readable knowledge substrate for coding agents. As AI-written commits lack semantic trails, institutional knowledge loss compounds run-over-run. Directly actionable: structured commits become persistent agent memory...
Skill-based architectures let agents generate, refine and reuse procedures from past runs, which means a malicious capability can be written into the skill store as a legitimate artifact (arXiv 2608.30429). SARGE red-teams that pipeline through iterative generation, escalation...
Skill-SP couples a proposer, solver and dynamic skill controller in an RL loop where each skill guarantees deep verifiable execution in a specific scenario while routing across skills preserves variety. Environment-bound methods get precise feedback but stay narrow; open-ended...
MindPattern daily
One email a day at 7 AM. Sources and a take on every story. Unsubscribe anytime.