Linux Blimp

Intelligent Email Agent v1.2 — Release Announcement:
Art2Dec SoftLab · mshell Ecosystem · May 2026

We are pleased to announce version 1.2 of the Intelligent Email Agent — the most complete autonomous email management system for IMAP/POP3 mail servers available today. Built on the mshell Ecosystem, it combines a native GTK3 GUI with a Go IMAP/POP3 client and a dual-LLM pipeline for classification, draft generation, and autonomous sending. The system is model-agnostic and works with Claude, GPT, and open-source models.

New in Version 1.2:
a) Bulk Contact Management
Checkbox selection in both the main contact list and Archive Browser, Shift+Click range selection for fast multi-contact operations, Bulk Archive, Restore and Restore Bulk operations, permanent deletion with single confirmation.
b) Remove Completely
c) Permanent deletion of archived contacts with full audit trail
d) Full Text Search
Deep search across all threads and attachments: PDF, DOCX, XLSX, PPTX, TXT, MD, HTML, TEX
Scoped separately for active contacts and archive
Extension search: by names or type .pdf, .doc/.docx, .md, .txt, .jpg, .jpeg, .mp4, .mpg, .mpeg and others to find all files of that type and searching content at non-binary files.
Double-click result opens thread or file at the exact match location. Intelligent Email Agent v1.2 · May 2026 · Art2Dec SoftLab (non-profit softlab)

What is Materialization at mshell Workflow?
Materialization is the process of converting a generative workflow into a deterministic, LLM-free version. When you run a generative .md file with MSHELL_MATERIALIZE=1, mshell creates a _materialized.md file alongside it.
The materialized file contains the actual code generated by the LLMs during that run — with all LLM directives replaced by the concrete code they produced. Running the materialized file executes the same logic every time, without any LLM calls.

Intelligent Email Agent for IMAP/POP3 Mail Systems, v.1.0, April 2026.

This benchmark measures execution time of identical computational tasks implemented natively in six languages
(Python, C, C++, Rust, Go, Lua) plus Bash system utilities (wc, grep, awk). The goal is to understand the real-
world performance characteristics of each language as a pipeline stage in mshell workflows, including pure
computation speed, system operation speed, pipeline overhead (including compilation), and repeatability across
runs.