--- title: "Azure AI Content Safety for Python: review, setup & limitations · Undominated.ai" canonical: https://undominated.ai/skills/microsoft-azure-ai-contentsafety-py/ description: "Pins Microsoft's Python Content Safety skill, which keeps new code on Entra credentials and can stop analysis when a blocklist matches." --- # Azure AI Content Safety for Python: review, setup & limitations · Undominated.ai > Pins Microsoft's Python Content Safety skill, which keeps new code on Entra credentials and can stop analysis when a blocklist matches. [← Explore all skills](/skills/) CONTENT MODERATION / Microsoft # Azure AI Content Safety for Python Pins Microsoft's Python Content Safety skill, which keeps new code on Entra credentials and can stop analysis when a blocklist matches. [See setup guidance ↓](#setup)[Original source ↗](https://github.com/microsoft/skills/tree/fc164ba1d7a0c10377dbcc1c9ac07253b9329fdb/.github/plugins/azure-sdk-python/skills/azure-ai-contentsafety-py) SOURCE REVIEW Reviewed 2026-10-15 Evidence 6 linked sources Publisher Microsoft Licence [MIT ↗](https://github.com/microsoft/skills/blob/fc164ba1d7a0c10377dbcc1c9ac07253b9329fdb/LICENSE) Revision fc164ba1d7a0 [Read what was—and wasn’t—checked ↓](#review) “Azure AI Content Safety SDK for Python. Use for detecting harmful content in text and images with multi-severity classification. Triggers: "azure-ai-contentsafety", "ContentSafetyClient", "content moderation", "harmful content", "text analysis", "image analysis". ” [Microsoft · upstream description ↗](https://github.com/microsoft/skills/blob/fc164ba1d7a0c10377dbcc1c9ac07253b9329fdb/.github/plugins/azure-sdk-python/skills/azure-ai-contentsafety-py/SKILL.md) Our analysis follows below. 01 / THE REASONING ## Why this made the selection. - New code is told to use DefaultAzureCredential, and AzureKeyCredential is limited to an existing keyed deployment that has not been migrated to Entra ID. - A text analysis can set halt_on_blocklist_hit so a blocklist match stops immediately, and the best-practice list says to use that for immediate rejection and to pre-moderate model output before it is shown. ### A good fit for - Moderating text or images when new code must use Entra credentials and a named blocklist match should stop the analysis ### Weigh up before choosing - This review did not execute the resource. - `/plugin install azure-sdk-python@skills` installs the azure-sdk-python plugin named in the plugin README, not this file alone. - `npx skills add microsoft/skills --skill azure-sdk-python` also names that plugin, not this file alone. - The repository README's `/plugin install azure@skills` installs a different plugin. - Blocklist create and item-update examples have no confirmation step and no dry-run. - The file does not choose a numeric severity cutoff. The first analyze example does not set halt_on_blocklist_hit. - references/capabilities.md and references/non-hero-scenarios.md were not opened. The skill frontmatter version is 1.0.0. 02 / THE REVIEW RECORD ## What we actually inspected. Source review has boundaries. A clear record is more useful than a “safe” badge. ### Material inspected - .github/plugins/azure-sdk-python/skills/azure-ai-contentsafety-py/SKILL.md - README.md - .github/plugins/azure-sdk-python/README.md - .github/plugins/azure-sdk-python/.claude-plugin/plugin.json - LICENSE ### Our findings - Production DefaultAzureCredential is constrained with AZURE_TOKEN_CREDENTIALS=prod, and API keys are avoided because they bypass Entra audit and rotation. - AzureKeyCredential is reserved for an existing keyed deployment that has not moved to Entra ID, and CONTENT_SAFETY_KEY is marked as only for that legacy path. - The blocklist analysis sample sets halt_on_blocklist_hit, and the best practices say to use it for immediate rejection. - The best practices say to pre-moderate AI outputs before showing them to users. - Sync and async Content Safety clients must not be mixed, and clients are closed with context managers so token caches are released. ### Not established by this review - The resource was not executed and its tools were not called. - No text or image was analyzed and no blocklist was created or updated. - The referenced capability and non-hero files were not opened. - The azure-sdk-python plugin was not installed. - Not executed: it was not installed or run. The review applies to the material and revision named here. A newer upstream release can change its behavior. 03 / PUT IT TO WORK ## Add a skill to your workflow. [Upstream setup instructions ↗](https://github.com/microsoft/skills/blob/fc164ba1d7a0c10377dbcc1c9ac07253b9329fdb/.github/plugins/azure-sdk-python/README.md) DOCUMENTED COMMAND /plugin install azure-sdk-python@skills Copy command ↗ Copying does not execute this command. It may retrieve a newer version than the reviewed source. - The azure-sdk-python plugin README shows `/plugin install azure-sdk-python@skills`. That installs the azure-sdk-python plugin, not this one file. - The same README shows `npx skills add microsoft/skills --skill azure-sdk-python`, which also names the plugin rather than this file. - The repository README shows `/plugin marketplace add microsoft/skills` and `/plugin install azure@skills` for a different plugin, plus `npx skills add microsoft/skills` for a repository wizard. ### Before you start - The skill shows `pip install azure-ai-contentsafety`. This review did not run it. - CONTENT_SAFETY_ENDPOINT as named in the skill - DefaultAzureCredential through Azure CLI, VS Code, Developer CLI, or a managed identity, as the skill describes - CONTENT_SAFETY_KEY only for the legacy API-key path the skill limits to an unmigrated keyed deployment - Copilot CLI for the slash-command install line ### Compatibility GitHub Copilot · GitHub Copilot CLI ### Implementation python ### Local skill text; Content Safety calls run only if an operator executes them - Read the skill text - Authenticate to Azure only if the operator uses a credential the skill names - Analyze text or images, or create and update a blocklist, only if the operator runs those client calls - The image example reads local image.jpg, base64-encodes its bytes and sends them in an analyze_image request; the URL example instead supplies a blob_url. ### Cost model MIT; no Azure or Python command was run 04 / FOLLOW THE EVIDENCE ## The source trail. Our notes are separate from the original resource. Check upstream before adopting a new version. - [azure-ai-contentsafety-py SKILL.md ↗](https://github.com/microsoft/skills/blob/fc164ba1d7a0c10377dbcc1c9ac07253b9329fdb/.github/plugins/azure-sdk-python/skills/azure-ai-contentsafety-py/SKILL.md) Checked 2026-10-14 https://github.com/microsoft/skills/blob/fc164ba1d7a0c10377dbcc1c9ac07253b9329fdb/.github/plugins/azure-sdk-python/skills/azure-ai-contentsafety-py/SKILL.md Supports: Entra credential preference, legacy key limit, blocklist halt, and pre-moderation - [azure-sdk-python plugin README ↗](https://github.com/microsoft/skills/blob/fc164ba1d7a0c10377dbcc1c9ac07253b9329fdb/.github/plugins/azure-sdk-python/README.md) Checked 2026-10-14 https://github.com/microsoft/skills/blob/fc164ba1d7a0c10377dbcc1c9ac07253b9329fdb/.github/plugins/azure-sdk-python/README.md Supports: /plugin install azure-sdk-python@skills and the npx plugin install line - [azure-sdk-python plugin.json ↗](https://github.com/microsoft/skills/blob/fc164ba1d7a0c10377dbcc1c9ac07253b9329fdb/.github/plugins/azure-sdk-python/.claude-plugin/plugin.json) Checked 2026-10-14 https://github.com/microsoft/skills/blob/fc164ba1d7a0c10377dbcc1c9ac07253b9329fdb/.github/plugins/azure-sdk-python/.claude-plugin/plugin.json Supports: plugin name azure-sdk-python and MIT - [Repository README ↗](https://github.com/microsoft/skills/blob/fc164ba1d7a0c10377dbcc1c9ac07253b9329fdb/README.md) Checked 2026-10-14 https://github.com/microsoft/skills/blob/fc164ba1d7a0c10377dbcc1c9ac07253b9329fdb/README.md Supports: MIT statement, Copilot CLI install, and a different azure plugin install line - [Repository LICENSE ↗](https://github.com/microsoft/skills/blob/fc164ba1d7a0c10377dbcc1c9ac07253b9329fdb/LICENSE) Checked 2026-10-14 https://github.com/microsoft/skills/blob/fc164ba1d7a0c10377dbcc1c9ac07253b9329fdb/LICENSE Supports: MIT License - [Pinned definition rechecked for the 2026-10-15 resource wave ↗](https://raw.githubusercontent.com/microsoft/skills/fc164ba1d7a0c10377dbcc1c9ac07253b9329fdb/.github/plugins/azure-sdk-python/skills/azure-ai-contentsafety-py/SKILL.md) Checked 2026-10-15 https://raw.githubusercontent.com/microsoft/skills/fc164ba1d7a0c10377dbcc1c9ac07253b9329fdb/.github/plugins/azure-sdk-python/skills/azure-ai-contentsafety-py/SKILL.md Supports: upstreamDescription, access.permissions, review KEEP COMPARING ## Other approaches to consider. Related by category or shared topics. These are alternatives to inspect, not a measured quality order. [### Azure AI Document Translation for Python ↗ Pins Microsoft's Python document-translation skill, which uses Entra for new clients and limits blob SAS permissions on source and target containers.](/skills/microsoft-azure-ai-translation-document-py/)[### Azure AI Language Conversations for Python ↗ Pins Microsoft's Python conversation-analysis skill, which tells new code to use token credentials and reserves API keys for unmigrated keyed deployments.](/skills/microsoft-azure-ai-language-conversations-py/)[### Azure AI Text Analytics for Python ↗ Pins Microsoft's Python text-analytics skill, which keeps new code on DefaultAzureCredential and reserves API keys for unmigrated keyed deployments.](/skills/microsoft-azure-ai-textanalytics-py/) [AI Tools ↗](/tools/)[Skills ↗](/skills/)[Agents ↗](/agents/)[MCP Servers ↗](/mcp-servers/)[Workflows ↗](/workflows/) ## Continue your investigation - [Choose a toolkit](/tools/) - [Inspect execution hosts](/agents/) - [Check tool access](/mcp-servers/) - [Check project dependencies](/check-repo/)