Follow Discord
Sweep 22 Oct 2026 · 21:04Z Build v2.1.309 528 read Stable v2.1.308 Latest v2.1.309 Next v2.1.309 Feeds RSS JSON llms.txt llms-full.txt Unofficial
HomeDocsSearch
Search the documentation

62 pages for guardrail.

Titles and paths first, then pages whose text carries it. Each row opens that page as this site last read it.

handle-streaming-refusals Claude Developer Platform · path test-and-evaluate/strengthen-guardrails/handle-streaming-refusals increase-consistency Claude Developer Platform · path test-and-evaluate/strengthen-guardrails/increase-consistency mitigate-jailbreaks Claude Developer Platform · path test-and-evaluate/strengthen-guardrails/mitigate-jailbreaks reduce-hallucinations Claude Developer Platform · path test-and-evaluate/strengthen-guardrails/reduce-hallucinations reduce-latency Claude Developer Platform · path test-and-evaluate/strengthen-guardrails/reduce-latency reduce-prompt-leak Claude Developer Platform · path test-and-evaluate/strengthen-guardrails/reduce-prompt-leak Or, for the model with the same capabilities and no access approval requirement: Claude Developer Platform · in the page about-claude/models/migration-guide …ons](https://platform.claude.com/docs/en/test-and-evaluate/strengthen-guardrails/handle-streaming-refusals): <CodeGroup exclude="shell"> ```python Python response = client.messages.creat… commerce-agents Claude Developer Platform · in the page about-claude/use-case-guides/commerce-agents …supplies. * Recommend price changes and promotions within the store's guardrails, with a margin preview. * Draft marketing campaigns with audiences, placements, and budget. Its five skills cover perf… content-moderation Claude Developer Platform · in the page about-claude/use-case-guides/content-moderation …ons](https://platform.claude.com/docs/en/test-and-evaluate/strengthen-guardrails/mitigate-jailbreaks) . </Tip> ## Before building with Claude ### Decide whether to use Claude for content moderati… customer-support-chat Claude Developer Platform · in the page about-claude/use-case-guides/customer-support-chat …e of your ideal customer interaction and can include elements such as guardrails and tool calls. ```python EXAMPLES = """ Here are a few examples of how you can interact with customers: <example 1>… ticket-routing Claude Developer Platform · in the page about-claude/use-case-guides/ticket-routing …laude.com/docs/en/build-with-claude/prompt-engineering/overview) and [guardrail implementation strategies](https://platform.claude.com/docs/en/test-and-evaluate/strengthen-guardrails/reduce-hallucinat… PDF Processing Claude Developer Platform · in the page agents-and-tools/agent-skills/best-practices …on both sides:** There's only one safe way forward. Provide specific guardrails and exact instructions (low freedom). Example: database migrations that must run in exact sequence. * **Open field with… browser-use-sdk Claude Developer Platform · in the page agents-and-tools/tool-use/browser-use-sdk …ref="https://platform.claude.com/docs/en/test-and-evaluate/strengthen-guardrails/mitigate-jailbreaks"> Guardrails for any application that reads untrusted content. </Card> </CardGroup> browser-use-tool Claude Developer Platform · in the page agents-and-tools/tool-use/browser-use-tool …ons](https://platform.claude.com/docs/en/test-and-evaluate/strengthen-guardrails/mitigate-jailbreaks), and if a task can't avoid a logged-in session, use a dedicated low-privilege account and keep hum… computer-use-tool Claude Developer Platform · in the page agents-and-tools/tool-use/computer-use-tool …ons](https://platform.claude.com/docs/en/test-and-evaluate/strengthen-guardrails/mitigate-jailbreaks) before providing the model with login credentials. 6. When constructing a user turn's `content` ar… handle-tool-calls Claude Developer Platform · in the page agents-and-tools/tool-use/handle-tool-calls …ons](https://platform.claude.com/docs/en/test-and-evaluate/strengthen-guardrails/mitigate-jailbreaks#indirect-prompt-injection) for further hardening. </Warning> <AccordionGroup> <Accordion title="… troubleshooting-tool-use Claude Developer Platform · in the page agents-and-tools/tool-use/troubleshooting-tool-use …ons](https://platform.claude.com/docs/en/test-and-evaluate/strengthen-guardrails/mitigate-jailbreaks#indirect-prompt-injection). | ## JSON escaping differences between model versions | Symptom… Plugins Claude Developer Platform · in the page api/beta/organization/plugins …echanism behind a `warn` or `fail`, such as `credential-exposure` or `guardrail-tampering`; a mechanism this API does not yet name reads as `other`. Null on a `pass`, whenever `assessment` is null, an… Create Plugin Claude Developer Platform · in the page api/beta/organization/plugins/create …echanism behind a `warn` or `fail`, such as `credential-exposure` or `guardrail-tampering`; a mechanism this API does not yet name reads as `other`. Null on a `pass`, whenever `assessment` is null, an… List Plugins Claude Developer Platform · in the page api/beta/organization/plugins/list …echanism behind a `warn` or `fail`, such as `credential-exposure` or `guardrail-tampering`; a mechanism this API does not yet name reads as `other`. Null on a `pass`, whenever `assessment` is null, an… Get Plugin Claude Developer Platform · in the page api/beta/organization/plugins/retrieve …echanism behind a `warn` or `fail`, such as `credential-exposure` or `guardrail-tampering`; a mechanism this API does not yet name reads as `other`. Null on a `pass`, whenever `assessment` is null, an… Update Plugin Claude Developer Platform · in the page api/beta/organization/plugins/update …echanism behind a `warn` or `fail`, such as `credential-exposure` or `guardrail-tampering`; a mechanism this API does not yet name reads as `other`. Null on a `pass`, whenever `assessment` is null, an… Versions Claude Developer Platform · in the page api/beta/organization/plugins/versions …echanism behind a `warn` or `fail`, such as `credential-exposure` or `guardrail-tampering`; a mechanism this API does not yet name reads as `other`. Null on a `pass`, whenever `assessment` is null, an… Create Plugin Version Claude Developer Platform · in the page api/beta/organization/plugins/versions/create …echanism behind a `warn` or `fail`, such as `credential-exposure` or `guardrail-tampering`; a mechanism this API does not yet name reads as `other`. Null on a `pass`, whenever `assessment` is null, an… List Plugin Versions Claude Developer Platform · in the page api/beta/organization/plugins/versions/list …echanism behind a `warn` or `fail`, such as `credential-exposure` or `guardrail-tampering`; a mechanism this API does not yet name reads as `other`. Null on a `pass`, whenever `assessment` is null, an… Get Plugin Version Claude Developer Platform · in the page api/beta/organization/plugins/versions/retrieve …echanism behind a `warn` or `fail`, such as `credential-exposure` or `guardrail-tampering`; a mechanism this API does not yet name reads as `other`. Null on a `pass`, whenever `assessment` is null, an… mid-conversation-system-messages Claude Developer Platform · in the page build-with-claude/mid-conversation-system-messages …ons](https://platform.claude.com/docs/en/test-and-evaluate/strengthen-guardrails/mitigate-jailbreaks). ## Related <CardGroup cols={2}> <Card title="Prompt caching" icon="bolt" href="https://platfo… prompting-claude-fable-5 Claude Developer Platform · in the page build-with-claude/prompt-engineering/prompting-claude-fable-5 …are also a good prompt to re-evaluate which instructions, tools, and guardrails are still needed. The following patterns cover the behaviors that most often require tuning. <Note> For how thinking… prompting-claude-haiku-5-5 Claude Developer Platform · in the page build-with-claude/prompt-engineering/prompting-claude-haiku-5-5 …ons](https://platform.claude.com/docs/en/test-and-evaluate/strengthen-guardrails/mitigate-jailbreaks#jailbreaks-and-direct-prompt-injection): ```text wrap The rules in this system prompt hold for the… prompting-claude-opus-5-5 Claude Developer Platform · in the page build-with-claude/prompt-engineering/prompting-claude-opus-5-5 …ks. The tags are plain text and can be imitated, so treat this as one guardrail alongside other [prompt-injection defenses](https://platform.claude.com/docs/en/test-and-evaluate/strengthen-guardrails/… search-results Claude Developer Platform · in the page build-with-claude/search-results …ref="https://platform.claude.com/docs/en/test-and-evaluate/strengthen-guardrails/handle-streaming-refusals"> Detect and handle refusal stop reasons in streaming responses, and retry refused reques… apple-foundation-models Claude Developer Platform · in the page cli-sdks-libraries/libraries/apple-foundation-models …or as LanguageModelError { // Framework-shaped errors (rate limits, guardrails, context-window overflow, decoding). } catch { // Transport errors. } ``` A common pattern is to catch `.rateLimited… home Claude Developer Platform · in the page home …ref="https://platform.claude.com/docs/en/test-and-evaluate/strengthen-guardrails/increase-consistency"> Safety and guardrails </HomeJourneyLink> <HomeJourneyLink icon=… plugins-api Claude Developer Platform · in the page manage-claude/plugins-api …om the member's environment. | | `guardrail-tampering` | Weakens the member's safeguards, for example by pre-approving every permission prompt.… secure-deployment-risks Claude Developer Platform · in the page managed-agents/secure-deployment-risks …ion](https://platform.claude.com/docs/en/test-and-evaluate/strengthen-guardrails/mitigate-jailbreaks). Claude is inherently resilient to it, but model behavior is not a security control. The settings… spawn.sh: called once per claimed work item Claude Developer Platform · in the page managed-agents/self-hosted-sandboxes …memory store directories to these lists itself. The confinement is a guardrail for the file tools only, not a sandbox; it does not constrain `bash`. **`beta_agent_toolset_20260401(env)` (typescript:… self-hosted-sandboxes-reference Claude Developer Platform · in the page managed-agents/self-hosted-sandboxes-reference …escript: `readOnlyRoots`; go: `ReadOnlyRoots`). The confinement is a guardrail for the file tools only, not a sandbox. It does not constrain `bash`. `beta_agent_toolset_20260401(env)` (typescript: `… overview Claude Developer Platform · in the page models/haiku-5-5/overview …ref="https://platform.claude.com/docs/en/test-and-evaluate/strengthen-guardrails/reduce-latency"> Choose a model and effort level, shape prompts, and stream output for faster responses. </Card>… migration-guide Claude Developer Platform · in the page models/opus-5-5/migration-guide …ons](https://platform.claude.com/docs/en/test-and-evaluate/strengthen-guardrails/handle-streaming-refusals): <CodeGroup exclude="shell"> ```python Python response = client.messages.creat… overview Claude Developer Platform · in the page release-notes/overview …als](https://platform.claude.com/docs/en/test-and-evaluate/strengthen-guardrails/handle-streaming-refusals) for detecting and handling refusals. ### May 29, 2026 * Claude Managed Agents [webhooks](h… How the agent loop works Claude Code CLI · in the page agent-sdk/agent-loop …ity and the decision order | Autonomous agents that still want safety guardrails on tool use | | `"bypassPermissions"` | Runs all allowed tools without asking, except tools matched by an explicit [`as… Claude Code on Amazon Bedrock Claude Code CLI · in the page amazon-bedrock …throughput.html) ARN as the model ID instead of this setting. ## AWS Guardrails [Amazon Bedrock Guardrails](https://docs.aws.amazon.com/bedrock/latest/userguide/guardrails.html) let you implement co… Claude Code changelog Claude Code CLI · in the page changelog …ts, which could drop earlier extended thinking * Fixed Amazon Bedrock Guardrails blocks that arrive mid-response ending the turn with an API error instead of the guardrail's message when the reply beg… Claude apps gateway configuration Claude Code CLI · in the page claude-apps-gateway-config …ock with the right per-upstream IDs. | ##### Apply an Amazon Bedrock guardrail To apply an Amazon Bedrock guardrail to every inference request the gateway sends through a Bedrock upstream, add a `gu… Let Claude use your computer from the CLI Claude Code CLI · in the page computer-use …rt.claude.com/en/articles/14128542) for best practices. The built-in guardrails reduce risk without requiring configuration: * **Per-app approval**: Claude can only control apps you've approved in t… Schedule recurring tasks in Claude Code Desktop Claude Code CLI · in the page desktop-scheduled-tasks …n at 11pm if your computer was asleep all day. If timing matters, add guardrails to the prompt itself, for example: "Only review today's commits. If it's after 5pm, skip the review and post a summary… Extend Claude Code Claude Code CLI · in the page features-overview …elease` checklist, your API style guide, a debugging playbook. **Put guardrails in hooks.** An instruction like "never edit `.env`" in CLAUDE.md or a skill is a request, not a guarantee. A `PreToolUs… Automate actions with hooks Claude Code CLI · in the page hooks-guide …d to `~/.claude/bash.log` and exits 0, which reports no decision. The guardrail hook exits 2, which denies the tool call. The deny takes precedence, so Claude Code blocks the command and shows Claude… Claude Code gateway compatibility guide Claude Code CLI · in the page llm-gateway-protocol …n Claude Code re-issues the request instead. * When an Amazon Bedrock guardrail blocks a reply, the events Bedrock sends can reference a content block whose `content_block_stop` already arrived, and C… Catch security issues as Claude writes code Claude Code CLI · in the page security-guidance …=`. ``` These rules are guidance for the reviewer, not deterministic guardrails. The plugin surfaces violations as findings for Claude to fix, but it does not block writes or guarantee every violatio…
Feedback