August 2026 builds

Atesor AI is a LangGraph multi-agent system that autonomously ports x86/ARM packages to RISC-V (riscv64). Each package below was cloned from upstream, built natively inside a riscv64 Docker sandbox (Alpine/musl or Debian/glibc), fixed where it broke, and packaged with a reproducible recipe.

Every *.zip asset contains the ported source tree, the porting build_recipe.md, and the per-package build logs. Assets are uploaded incrementally as CI jobs (and the nightly retry workflow) complete throughout August 2026 — no need to wait for a single "green" run.

Filename format: <package>-<YYYYMMDD>-<HHMMSS>-<alpine|debian>.zip.

GitHub caps each release at 1000 assets, so once a monthly release fills up the workflow rolls over to builds-2026-08-NN (August 2026 builds NN) — all overflow releases for the same month list the same packages, just split across releases.

July 2026 builds 01

Atesor AI is a LangGraph multi-agent system that autonomously ports x86/ARM packages to RISC-V (riscv64). Each package below was cloned from upstream, built natively inside a riscv64 Docker sandbox (Alpine/musl or Debian/glibc), fixed where it broke, and packaged with a reproducible recipe.

Every *.zip asset contains the ported source tree, the porting build_recipe.md, and the per-package build logs. Assets are uploaded incrementally as CI jobs (and the nightly retry workflow) complete throughout July 2026 — no need to wait for a single "green" run.

Filename format: <package>-<YYYYMMDD>-<HHMMSS>-<alpine|debian>.zip.

GitHub caps each release at 1000 assets, so once a monthly release fills up the workflow rolls over to builds-2026-07-NN (July 2026 builds NN) — all overflow releases for the same month list the same packages, just split across releases.

July 2026 builds

Atesor AI is a LangGraph multi-agent system that autonomously ports x86/ARM packages to RISC-V (riscv64). Each package below was cloned from upstream, built natively inside a riscv64 Docker sandbox (Alpine/musl or Debian/glibc), fixed where it broke, and packaged with a reproducible recipe.

Every *.zip asset contains the ported source tree, the porting build_recipe.md, and the per-package build logs. Assets are uploaded incrementally as CI jobs (and the nightly retry workflow) complete throughout July 2026 — no need to wait for a single "green" run.

Filename format: <package>-<YYYYMMDD>-<HHMMSS>-<alpine|debian>.zip.

GitHub caps each release at 1000 assets, so once a monthly release fills up the workflow rolls over to builds-2026-07-NN (July 2026 builds NN) — all overflow releases for the same month list the same packages, just split across releases.

June 2026 builds 02

Atesor AI is a LangGraph multi-agent system that autonomously ports x86/ARM packages to RISC-V (riscv64). Each package below was cloned from upstream, built natively inside a riscv64 Docker sandbox (Alpine/musl or Debian/glibc), fixed where it broke, and packaged with a reproducible recipe.

Every *.zip asset contains the ported source tree, the porting build_recipe.md, and the per-package build logs. Assets are uploaded incrementally as CI jobs (and the nightly retry workflow) complete throughout June 2026 — no need to wait for a single "green" run.

Filename format: <package>-<YYYYMMDD>-<HHMMSS>-<alpine|debian>.zip.

GitHub caps each release at 1000 assets, so once a monthly release fills up the workflow rolls over to builds-2026-06-NN (June 2026 builds NN) — all overflow releases for the same month list the same packages, just split across releases.

June 2026 builds 01

Atesor AI is a LangGraph multi-agent system that autonomously ports x86/ARM packages to RISC-V (riscv64). Each package below was cloned from upstream, built natively inside a riscv64 Docker sandbox (Alpine/musl or Debian/glibc), fixed where it broke, and packaged with a reproducible recipe.

Every *.zip asset contains the ported source tree, the porting build_recipe.md, and the per-package build logs. Assets are uploaded incrementally as CI jobs (and the nightly retry workflow) complete throughout June 2026 — no need to wait for a single "green" run.

Filename format: <package>-<YYYYMMDD>-<HHMMSS>-<alpine|debian>.zip.

GitHub caps each release at 1000 assets, so once a monthly release fills up the workflow rolls over to builds-2026-06-NN (June 2026 builds NN) — all overflow releases for the same month list the same packages, just split across releases.

atesor-v1.2.0

Setup

Review the project's README.md for details.

What's Changed

• Ci dev by @akifejaz in #9
• Add CI support and adash release dashboard by @akifejaz in #10

Full Changelog: builds-2026-06...beta

June 2026 builds

Atesor AI is a LangGraph multi-agent system that autonomously ports x86/ARM packages to RISC-V (riscv64). Each package below was cloned from upstream, built natively inside a riscv64 Docker sandbox (Alpine/musl or Debian/glibc), fixed where it broke, and packaged with a reproducible recipe.

Every *.zip asset contains the ported source tree, the porting build_recipe.md, and the per-package build logs. Assets are uploaded incrementally as CI jobs (and the nightly retry workflow) complete throughout June 2026 — no need to wait for a single "green" run.

Filename format: <package>-<YYYYMMDD>-<HHMMSS>-<alpine|debian>.zip.

atesor-v1.0.0

Quick Start

Prerequisites

• Docker installed and running.
• Cross-Platform Support: If you are on an x86 host, you must enable RISC-V emulation via binfmt:

  docker run --privileged --rm tonistiigi/binfmt --install all

• API key for an LLM provider (OpenAI, Gemini, or OpenRouter).

Installation

# Clone the repository
git clone https://github.com/akifejaz/atesor-ai
cd atesor-ai

# Install dependencies
pip install -r requirements.txt

# Set up environment variables
cp .env-example .env
# Edit .env and add your API keys

Basic Usage

# 1. Prepare the RISC-V Sandbox
python3 main.py --setup-only

# 2. Start Porting a Package
python3 main.py --repo https://github.com/madler/zlib --verbose

# 3. Force a clean rebuild with custom attempt limit
python3 main.py --repo https://github.com/madler/zlib --force --max-attempts 8

Flag	Default	Description
--repo URL	required	GitHub repository URL to port
--verbose	false	Enable DEBUG logging to console
--setup-only	false	Initialize sandbox without porting
--max-attempts	5	Maximum fix attempts before escalation
--force	false	Force fresh clone and rebuild

---

Environment Variables

Variable	Purpose
LLM_PROVIDER	gemini (default), openai, or openrouter
GOOGLE_API_KEY	Required for Gemini
OPENAI_API_KEY	Required for OpenAI
OPENROUTER_API_KEY	Required for OpenRouter
LANGCHAIN_API_KEY	Optional — LangSmith tracing
LANGCHAIN_TRACING_V2	Set to true to activate tracing

Output Locations

Path	Content
workspace/output/{repo}_report_*.md	Markdown porting guide
workspace/output/{repo}_state_*.json	Full state snapshot
workspace/logs/agent.log	DEBUG-level agent log
workspace/logs/agent-call.log	LLM call audit trail

---

Project Structure

• main.py: Entry point for CLI and Docker management.
• src/graph.py: The core LangGraph state machine (all agent nodes + routing).
• src/scripted_ops.py: Zero-cost analysis and repo management.
• src/state.py: Global process tracking, error classification, and data structures.
• src/tools.py: Safe command execution and file utilities.
• src/memory.py: Few-shot learning system with auto-learning.
• src/config.py: Environment-aware workspace path resolution.
• src/models.py: LLM provider factory and per-role model configuration.
• src/knowledge.py: Static RISC-V / Alpine knowledge base.
• src/artifact_scanner.py: Post-build artifact detection and RISC-V verification.
• src/llm_logger.py: LLM call audit trail logging.
• data/examples/: Curated few-shot examples per agent type.
• data/recipe_cache.json: Cache of successfully ported package recipes.
• tests/: Automated unit tests for engine logic.

---