Public story · 2026-03-03 · source-backed

AgentSkillOS: Skill Orchestration at Ecosystem Scale

Issue

2026-03-03

Confidence

source-backed

Sources

Redaction

passed

Story

First principled framework for selecting and orchestrating 200-200K skills via capability tree and DAG-based pipelines. Code released. Directly actionable for multi-skill agent systems. arXiv 2603.02176

Airbnb says AI writes 60% of its new code and cut concept-to-launch by 60%
Four stories about things going wrong. Here's one about something working, with actual numbers attached. In an August 7 disclosure covered by TechCrunch, Airbnb said AI now writes 60% of its new code, that concept-to-launch time on key initiatives has dropped by as much as 60%...
RSIBench-Data: agents improve on their first attempt 58% of the time but finish below their own peak in 78% of runs.
arXiv 2607.25886 isolates data-centric research capability by fixing the entire post-training stack so only the agent's data strategy varies. Four frontier agents across six benchmarks. Among searches that continued past the best observed score, 78.26% ended on a lower-scoring...
SplitAgent: Privacy-Preserving Enterprise Agents
Addresses the fundamental privacy dilemma: cloud models need data access but enterprises can't share sensitive information. Splits execution between enterprise-side privacy agents and cloud-side capability agents. Directly relevant to AWS AgentCore and enterprise adoption. arX...
FlashOptim: 50% Training Memory Reduction (Databricks)
Reduces per-parameter training memory by 50%+ through improved master weight splitting and 8-bit optimizer state quantization. AdamW drops from 16 bytes to 7 bytes per parameter. Tested on Llama-3.1-8B finetuning with no quality degradation. Code at databricks/flashoptim. Dire...
Adaptive skill libraries lift a fixed computer-use stack 5.7 to 18.6 points, conditionally.
An online skill-evolution framework turns interaction trajectories and evaluator feedback into a persistent versioned library, with each iteration executing against a frozen snapshot so evidence-guided updates only reach later iterations and no model parameters change (arXiv 2...
Score every skill edit by rolling the task back and re-running it.
Skill-α (arXiv 2608.01678) reframes skill generation as RL over sequential edits, decomposing skill construction into individually evaluable changes. The novel signal is a rollback reward that scores each modification by comparing downstream task execution using the original s...

Source trail

arXiv 2603.02176 More from arxiv.org

Entities

AgentSkillOS Code DAG Directly Ecosystem Scale First Skill Orchestration

Provenance

Canonical issue: Ramsay Research Agent — 2026-03-03
AI generated: no
Story unit: 2026-03-03-agentskillos-skill-orchestration-at-ecosystem-scale
Labels: source-backed, canonical briefing excerpt

Fetching from the wire…

AgentSkillOS: Skill Orchestration at Ecosystem Scale

Story

Related stories

Airbnb says AI writes 60% of its new code and cut concept-to-launch by 60%

RSIBench-Data: agents improve on their first attempt 58% of the time but finish below their own peak in 78% of runs.

SplitAgent: Privacy-Preserving Enterprise Agents

FlashOptim: 50% Training Memory Reduction (Databricks)

Adaptive skill libraries lift a fixed computer-use stack 5.7 to 18.6 points, conditionally.

Score every skill edit by rolling the task back and re-running it.

Source trail

Entities

Provenance

AgentSkillOS: Skill Orchestration at Ecosystem Scale

Story

Related stories

Airbnb says AI writes 60% of its new code and cut concept-to-launch by 60%

RSIBench-Data: agents improve on their first attempt 58% of the time but finish below their own peak in 78% of runs.

SplitAgent: Privacy-Preserving Enterprise Agents

FlashOptim: 50% Training Memory Reduction (Databricks)

Adaptive skill libraries lift a fixed computer-use stack 5.7 to 18.6 points, conditionally.

Score every skill edit by rolling the task back and re-running it.

Source trail

Entities

Provenance

Story

Related stories

Airbnb says AI writes 60% of its new code and cut concept-to-launch by 60%

RSIBench-Data: agents improve on their first attempt 58% of the time but finish below their own peak in 78% of runs.

SplitAgent: Privacy-Preserving Enterprise Agents

FlashOptim: 50% Training Memory Reduction (Databricks)

Adaptive skill libraries lift a fixed computer-use stack 5.7 to 18.6 points, conditionally.

Score every skill edit by rolling the task back and re-running it.

Tomorrow's issue, in your inbox.

Source trail

Entities

Provenance

Story

Related stories

Airbnb says AI writes 60% of its new code and cut concept-to-launch by 60%

RSIBench-Data: agents improve on their first attempt 58% of the time but finish below their own peak in 78% of runs.

SplitAgent: Privacy-Preserving Enterprise Agents

FlashOptim: 50% Training Memory Reduction (Databricks)

Adaptive skill libraries lift a fixed computer-use stack 5.7 to 18.6 points, conditionally.

Score every skill edit by rolling the task back and re-running it.

Tomorrow's issue, in your inbox.

Source trail

Entities

Provenance