#skills
8 posts tagged with "skills".
-
The Agent's Action Space Became a Training Variable
• 5 min readJune 2026 research stopped treating an agent's skills and tools as a fixed library and started co-optimizing them with the policy via RL.
-
Agents Are Learning the Memory Policy You Used to Hand-Code
• 5 min readA June 2026 wave moves the store/evict/retrieve decision from heuristics to a trained policy, and pushes consolidation into an offline sleep phase.
-
The Skill Supply Chain Got Poisoned Before It Got Secured
• 5 min readAgent skills are an executable supply chain that runs with your agent's full privileges — and the first wave of benchmarks shows our defenses see only half the attacks.
-
Optimization Is Moving From Weights to English
• 5 min readRecent work turns skills, harnesses, and context into objects you can search over and benchmark — optimizing the English around a frozen model instead of the model.
-
The Harness Got a Name
• 5 min readA new survey and Microsoft's BUILD 2026 release both landed on the same idea: agent capability is leaving the model and moving into the harness.
-
Skills Are the New SDK
• 5 min readOpenAI killed its visual Agent Builder the same week Google shipped first-party skills. The agent capability layer just consolidated.
-
The Agent Is a Workload, Not a Script
• 6 min readMid-May 2026 quietly shipped the operations layer for agents — versioned environments, runtime drain, behavior-based evals, portable skills.
-
The Agent Trust Stack Just Got Built: Three Weeks in May 2026
• 6 min readSkill cards, self-hosted sandboxes, MCP tunnels, computer-use verifiers, and a Five Eyes warning all landed in twenty-one days. The boring perimeter around capable agents finally has shape.