Capabilities · Concept
Agent Skill Supply Chains (ASSCs) and Dependency Risk
SkillConsist work sharpens dependency and consistency risks when reusable agent skills mix declarations, instructions, and code resources.
Forecast horizon
Connections
Connections · 3
How this node ties into the rest of the map, and the evidence behind each link.
from · informs 4/10
Cross-Task Skill Transfer Utility in LLM Agents Results show induced skills can harm transfer; subtask-level text skills and a utility score predict safer reuse.
+4 growth from · enables 3/10
MidTool: Mid-Training Data Synthesis for Agentic Tool Use Mid-training on tool APIs and MCP skills builds foundational agent tool-use competence before post-training.
+3 growth to · informs 2/10
AI Agents Reshaping Knowledge Work Agent skill supply chain analysis reveals governance gaps in how LLM agent capabilities are packaged and deployed for knowledge work.
+2 growthSignal sources
Signal sources
Dated facts from primary sources in this direction.
Task horizon doubling Mar 2025
The length of software tasks AI agents can do autonomously at 50% reliability has doubled about every 7 months — and since 2024 closer to every ~3 months.
METR → Benchmarks saturating Apr 2025
In one year scores rose by 18.8, 48.9 and 67.3 points on MMMU, GPQA and SWE-bench; real-world software solve rate jumped from 4.4% to 71.7%.
Stanford HAI — AI Index 2025 → Autonomous coding 2025–2026
On SWE-bench Verified (500 real GitHub issues), autonomous coding agents reached ~80–86% by late 2025, up from under 50% in early 2025.
Epoch AI →