CORTEX: A Verified Experience Layer for Generalist Agents

Garapati Keerthana; Manik Gupta

cs.AISep 27, 2026

CORTEX: A Verified Experience Layer for Generalist Agents

Authors: Garapati Keerthana, Manik Gupta

Organizations: Birla Institute of Technology and Science, Pilani, Hyderabad, India

Abstract

An agent can solve a task today and face the same task under new facts, tools, or governing knowledge tomorrow. Most agent systems can retrieve relevant text or recall prior conversations, but they lack a principled way to decide when a previous solution is still valid, when it must be adapted, and when it should be discarded. We introduce CORTEX (Contextual Orchestration and Reuse of Task EXperience), a general AI systems framework that connects specialized agents through an external layer of verified experience. Each episode records its task conditions, source and tool state, decisive predicates, proof trace, verifier, and outcome. A meta-controller chooses exact replay, checked adaptation, fresh synthesis, or escalation. Accepted episodes can become task patterns and procedural strategies through a challenge-driven development loop. This gives the system an implicit competence layer that can grow without changing model weights. We formalize system contracts for exact replay and source-version separation, and derive when reuse saves computation. A controlled two-domain implementation tests the exact-replay core on 1,000 synthetic cases. Complete-family holdouts test procedural transfer on 1,000 new-family cases across eight clinical and policy splits, with complete fresh-evidence grounding and perfect invariance to irrelevant-field and insertion-order perturbations. The transfer trace exposes the work required for verified strategy execution. These results establish an initial path toward general intelligence through reusable procedures, typed experience, and developmental transfer.

Figures & tables

Research line	Contribution and distinction
RAG	Grounds a current answer in retrieved text. CORTEX also checks whether an earlier derivation may govern a new task.
Agent memory	Preserves episodes or skills. CORTEX binds operational reuse to source, time, predicates, and a verifier.
Case reasoning	Supplies the retrieve–revise cycle. CORTEX makes replay and invalidation contracts explicit under changing authority.
Provenance	Tracks dependencies. CORTEX uses them in a task-level meta-controller and development loop.
Modular agents	Supply tools, roles, and realistic tests. CORTEX specifies the persistent competence layer connecting runs.

Table 1: Comparison by primary mechanism. The table is an architectural map, not an empirical ranking of prior systems.

Measure	Clinical	Policy
Cases / rule families	$500/5$	$500/3$
Official source artifacts	$9$	$4$
Exact match, experience core	$500/500$	$500/500$
Exact match, full evaluator	$500/500$	$500/500$
Current-edition snapshot	$251/500$	$249/500$
Edition-wide document diff	$46/500$	$0/500$

Table 2: Internal fidelity on controlled, source-curated cases. The scenario oracle and runtime share source interpretation, so exact agreement is not external expert validation. Grouped source-context review is not individual adjudication.

Measure	Fresh evidence	CORTEX strategy
New-family cases	$100$	$100$
Verified task completions	$100$	$100$
Full strategy execution	$100$	$100$
Fresh-evidence grounding	$100$	$100$
Mean reasoning work	$45.00$	$56.00$
Median response time (ms)	$10.14$	$10.82$

Table 3: New-family transfer. The strategy condition carries the generic procedure from four families and performs fresh source-bound evaluation on the fifth. Equal labels quantify internal fidelity under controlled source interpretation. Strategy bookkeeping exposes the cost of verified transfer.

Measure	All eight family splits
New-family cases	$1{,}000$
Family splits	$8$
Verified task completions	$1{,}000$
Full strategy executions	$1{,}000$
Fresh-evidence grounding	$1{,}000$
Input-invariance checks	$1{,}000$

Table 4: Cross-domain family transfer across the clinical and policy domains. The strategy condition preserves the generic procedure and performs fresh evaluation for every new family. Input-invariance checks compare results under irrelevant-field and insertion-order perturbations.

Condition	Correct	Verified completion
Direct agent	$4/4$	$100\%$
Typed roles	$4/4$	$100\%$
Typed roles + experience	$4/4$	$100\%$

Table 5: Logged narrow-agent pilot on four new-family cases. Equal scores show that direct and typed routes can all complete the transfer task under the shared protocol.

Explore similar work

KUPAS MASTER: Distilling the Tacit Expertise of Master Practitioners into Agent-Ready Experience Corpora
Sep 29, 2026Changmian Wang, Yuchao Ma, Xuchao Lu +13Expert Demonstrations Text Corpora
Shanghai Kupas Technology Co., Ltd. and Tongji University
Experience Compression Spectrum: Unifying Memory, Skills, and Rules in LLM Agents
Apr 17, 2026Xing Zhang, Guanghui Wang, Yanwei Cui +4Large Language Model Agents Long-Term Agent Memory
AWS Generative AI Innovation Center · HSBC Holdings Plc., HSBC Technology Center, China.
PreAct: Computer-Using Agents that Get Faster on Repeated Tasks
Jun 16, 2026Bojie LiComputer-Use Agents Replay
Pine AI

Sep 29, 2026, cs.AIJ/K move · Enter open · S save

KUPAS MASTER: Distilling the Tacit Expertise of Master Practitioners into Agent-Ready Experience Corpora

Changmian Wang, Yuchao Ma, Xuchao Lu, Chen Zhang+12

Shanghai Kupas Technology Co., Ltd. and Tongji University

Expert Demonstrations Text Corpora

Open paper AskPDF arXiv

Experienced professionals know more than just facts and conclusions. They know which cues matter, why a judgment is reasonable, and which action to take. Routine work records often leave out this tacit knowledge, making it difficult for Large Language Model (LLM) agents to use professional experience effectively. We introduce KUPAS MASTER, an experience engineering platform built around nine-layer cognitive corpus construction. It turns heterogeneous work records and practitioner interviews into traceable, reusable experience corpora for agents. Six case elements preserve the task process: context, cues, judgment, action, boundaries, and outcomes. Nine-layer cognitive corpus construction organizes tacit experience along nine extraction dimensions and stores the resulting assets in six libraries: rules, constraints, best practices, negative examples, corner cases, and skills. Semantic alignment, individual experience distillation, organizational consolidation, and cross-review preserve source evidence, conditions of use, and unresolved disagreements. The platform packages these assets into callable skills with explicit inputs, steps, dependencies, and stopping conditions, connecting experience collection to task execution and evaluation feedback. Using authorized samples from 20 randomly selected practitioners, the platform processed 1,576 source files into 23,024 individual experience records and 13,113 organizational assets. The evaluation spans multiple professional domains. Under common task inputs and scoring criteria, the base model, raw corpus retrieval-augmented generation (RAG), and KUPAS MASTER agent scored 70.63, 79.75, and 89.58, respectively. The KUPAS MASTER agent improved on raw-corpus RAG in all seven scoring dimensions. The platform provides a practical path from individual tacit experience to organizational knowledge and agent capabilities.

Abstract

Explore similar work

KUPAS MASTER: Distilling the Tacit Expertise of Master Practitioners into Agent-Ready Experience Corpora

Experience Compression Spectrum: Unifying Memory, Skills, and Rules in LLM Agents

PreAct: Computer-Using Agents that Get Faster on Repeated Tasks