Agents35d ago
UIUC Proposes Meta-Skills for Agent Harness Design in Test-Time AI4AI: 12% Performance Gain Without Weight Updates
Agent efficacy relies heavily on execution environments (harnesses) in addition to reasoning capability. UIUC researchers led by Heng Ji propose a test-time AI-for-AI (AI4AI) framework where a Builder agent learns 'Meta-Skills' to construct custom execution environments for a frozen Target agent. Learned from execution feedback, these principles guide when and what support resources to provide. Across Harness-Bench and NewtonBench, meta-skill harnesses improve macro-average performance by 8.95 percentage points over baseline and 12.02 points over direct skill delivery, paving the way for autonomous system-level self-improvement.
Agents26d ago
Skill Constellations: The First Supply Chain Audit of 2.19M GitHub Agent Skills Reveals Vulnerability Propagation Patterns
Modern AI coding agents such as Claude Code, Codex, and terminal assistants routinely execute agent skills—instructions and executable scripts bundled in SKILL.md specifications—with full local user privileges. Developers share these skills primarily by copying files across repositories, inadvertently forging a shadow software supply chain without centralized registries, semver tagging, or provenance tracking. A seminal empirical audit by Fahd Seddik (arXiv:2610.11169) introduces Skill Constellations, constructing the first dated copy network of agent skills derived from git provenance across 2,193,119 skill adoptions on GitHub. The investigation demonstrates that a minuscule set of root repositories generates nearly all downstream copies, yet GitHub star counts correlate poorly with true lineage sources. Critically, downstream copies almost never adopt upstream security patches. By fitting a predictive copy network model, auditing just the top 100 identified root repositories halts 14.9% of subsequent high-risk skill adoptions (compared to a negligible 0.5% when inspecting the top 100 starred repos), proving the urgent necessity for decentralized package ecosystems to adopt versioned references over loose copies.
Tooling5d ago
DeepSeek Harness team open-sources dsh-superagent, a proactive coding agent prototype for a single GitHub repo, alongside its Super-agent architecture article
On Oct 31, deepseek-ai open-sourced dsh-superagent-prototype on GitHub under the MIT license: a proactive coding agent that runs as a DeepSeek Harness bundle, lives on one GitHub repository and can optionally listen to Feishu. When an allowed person asks for work, it opens a work item, works in its own session and directory, and answers in the original conversation, for example with a pull request. The README states it is a research prototype, not an official DeepSeek product. The companion article, Exploring Super-agent Design, proposes a two-layer Router-Worker architecture; the team test, as reproduced by media, reports a p95 completion time of 15.7 minutes versus 29.5 minutes for a single-session setup.
Discussion & Comments
0Sign in to join the discussion
Connect with AI developers to exchange benchmark insights.