Full CIBA debug output, delete starter, sanitized binding message

Demo mode remediation returns success for CIBA flow

Add Outlook demo emails

Replace What is Amanat starter with Outlook scan

Add Remediate profile with note that destructive actions blocked in demo

Expand chat profile description with full demo context

Expand chainlit welcome: full demo context, repo link, watsonx explanation

Richer profile description, Kokota attribution

Remove welcome message, use chainlit.md for starters page with full context

Show welcome message in chat for demo mode

Add full demo description to HF Space page

Add auditable data link and watsonx context to welcome

CIBA step-up auth for destructive actions

Add model/timing footer to responses, explain demo context in welcome

Fix: parse tool_args from string when watsonx returns unparsed JSON

Show full traceback

Fresh IAM token per request, fix token expiry causing str.get error

Pin strands==1.34.1, use python:3.13-slim to match local

Show traceback in error message for debugging

Fix after_tool result extraction: handle str results from watsonx

Fix message extraction: handle str/dict/list response formats from watsonx

Switch to watsonx backend: Granite 4 H-Small on GPU, no local model needed

Revert to working python:3.12-slim + ollama CPU build

Use system python3 from ollama base, no version pinning

Use ollama base image directly with Python installed on top for GPU

Minimal Dockerfile: python:3.12-slim + ollama binary, cpu-upgrade

python:3.12-slim + CUDA runtime libs via apt (no deadsnakes, no nvidia base)

Fix: ubuntu22.04 not 24.04

Base on nvidia/cuda runtime for actual GPU access in ollama

Force NVIDIA_VISIBLE_DEVICES=all in entrypoint to override HF void setting

Add GPU diagnostics to startup logs

Copy ollama CUDA runners for GPU acceleration on T4

Copy ollama binary into python:3.13-slim — simplest approach

Multi-stage: Python 3.13 deps + Ollama CUDA runtime

Pull model at runtime instead of build time

Fix: wait for Ollama to start before pulling model

Switch to Ollama with GPU: pre-pulled granite4:micro, T4 tier

Revert to CPU build on cpu-upgrade — CUDA build stuck

Use pre-built CUDA binary instead of compiling from source

CUDA build + GPU offload for T4 tier

Disable GGML_NATIVE to avoid illegal instruction on different CPU

Increase ctx-size to 4096 — tool defs need 1500 tokens

Revert to granite-4.0-micro (dense) — hybrid arch not supported in llama.cpp b5580

Switch to granite-4.0-h-micro (3B hybrid) — h-small OOM on cpu-basic

Upgrade to Granite 4 H-Small (8B) for better tool calling

Force agent to call tools immediately

Add --jinja flag for tool calling support

Deploy Amanat: Granite 4 Micro + llama-server + Chainlit + Auth0

initial commit