RunningTab Adds Environment-Side Task Tracking to LLM Agents' Direct Workspace Interaction
kaist-ai · hf · 2026-10-08
KAIST presents RunningTab, a framework that fixes task-tracking gaps when LLM agents do Direct Workspace Interaction (DWI) — searching and reading workspace files from a terminal with no indexing.
- Problem: task requirements, files read, and files listed but never opened all slip through the context window, so agents can deliver reports missing key figures.
- Solution: an environment-side per-task tab where the agent registers requirements, the environment logs read files as provenance-tracked excerpts and unopened files as candidates; the agent resolves or explicitly sets aside each requirement, and a finish check returns still-open ones.
- Validated on three benchmarks with three LLMs, consistently outperforming plain DWI and model-side record baselines.
Related event: KAIST's RunningTab Tackles LLM Agent Long-Task Memory Loss(2 posts)→
More from coding & agent
- Ramp's hidden markdown-file offer read only by AI actually worked, customers claimed it — gaganghotra_ · 2026-10-08
- Creator revives a dead Notion course by feeding it to Codex as an interactive app — RichardsonDx · 2026-10-08
- Gary Bernhardt: AI 'factories' repeat the microservices over-engineering mistake — sergeykarayev · 2026-10-08
- Hierarchical RL with mixed discount rates may unlock long-horizon agent tasks — jessi_cata · 2026-10-08
- Graph-MIND: local MCP memory server hits 88.8% LongMemEval with zero LLM calls at write time — BrilliantGrocery8233 · 2026-10-08
- dotey's Claude Code workflow: Fable as Tech Lead delegating and reviewing for Opus — dotey · 2026-10-08