Azure AI Content Safety (Prompt Shields) vs Granite Guardian

Azure AI Content Safety (Prompt Shields)

Prompt Shields checks up to five retrieved documents for indirect injection, not only the user prompt. Needs an Azure subscription with a card, a resource and a region that has the feature, before the first call.

Granite Guardian

One ungated Apache 2.0 model judges harm, jailbreaks, RAG groundedness, function-call errors and custom criteria, with signed weights and published evaluation code. It is trained and tested on English only, each call checks one criterion, the 4.1 prompt format differs from 3.x, and IBM's watsonx.ai lists only the deprecated 3.0 model.

Azure AI Content Safety (Prompt Shields)

1. Send retrieved pages and tool results in the documents array of shieldPrompt, not in userPrompt, so document attacks are reported separately
2. Call text:shieldPrompt over REST with api-version=2024-09-01. The Python SDK 1.0.0 has no method for it
3. Keep each request under 10,000 characters across prompt and documents, and split long tool results
4. Create the resource in a region that lists Prompt Shields, since not every region has it
5. On F0 you get 5 requests a second. Queue checks or move to S0 before load testing

Granite Guardian

1. Append the <guardian> block as the final user message, with the mode line, ### Criteria: and ### Scoring Schema:. Copy the strings from the model card, because no package builds them
2. Use the no-think instruction for gating and parse <score>. Think mode writes a reasoning trace first, and the card's examples allow up to 2,048 output tokens
3. Treat yes as the criterion being met, which for built-in criteria means the risk is present. Treat a missing <score> tag as a failed check
4. Pass retrieved text through documents= and tool schemas through available_tools= in apply_chat_template, not inside the message text
5. Under Ollama, set num_ctx in the request options. IBM's docs say the default context is short and long requests are truncated