# Straza docs > Open source runtime governance for AI agents. These docs cover deployment, identity, policy, approvals, the MCP gateway, audit, security and the CLI. ## Evaluate - [Evaluate](https://docs.straza.ai/evaluate/index.md): What Straza is, how it works and where its limits are, then how to try it on a laptop or the demo stack. - [Decide whether Straza fits](https://docs.straza.ai/evaluate/decide-whether-straza-fits/index.md): The pages an IAM or security architect reads, in order, to decide whether to run a pilot. - [What Straza is](https://docs.straza.ai/evaluate/what-straza-is/index.md): What Straza does, where it decides, who it is for, and what it does not try to do. - [How it works](https://docs.straza.ai/evaluate/how-it-works/index.md): The parts of Straza, and how Straza decides each tool call before the tool runs. - [The security model](https://docs.straza.ai/evaluate/security-model/index.md): What Straza enforces and where, what it trusts, and which actions it does not see. - [What each lane stops](https://docs.straza.ai/evaluate/what-each-lane-stops/index.md): Read this before you test or review Straza: which lanes are a security boundary, what counts as a finding, and how to test each one. - [Identity and your IGA](https://docs.straza.ai/evaluate/identity-and-your-iga/index.md): What your identity provider, your IGA and Straza each decide, and how the first admin signs in. - [Try it on the demo stack](https://docs.straza.ai/evaluate/try-it-on-the-demo-stack/index.md): Run Straza with midPoint, Keycloak, AI agents and MCP servers on one machine, and see what the seeded stack does. - [Your first approval on one laptop](https://docs.straza.ai/evaluate/first-approval-on-one-laptop/index.md): Hold your agent's command on your laptop until you approve it. - [Try it on one laptop](https://docs.straza.ai/evaluate/try-it-on-one-laptop/index.md): Run Straza and Claude Code on your own computer. A starter rule denies a command, and you approve a held command in the browser. ## Concepts - [Concepts](https://docs.straza.ai/concepts/index.md): Each page explains a part of the Straza model and opens with the diagram from How it works, with that part highlighted. - [Roles and access](https://docs.straza.ai/concepts/roles-and-access/index.md): What a role lets its holders do, and who makes and assigns roles. - [Objects](https://docs.straza.ai/concepts/objects/index.md): How a role can be limited to some of a server's objects, on any server that declares its objects. - [Policies](https://docs.straza.ai/concepts/policy-model/index.md): How a rule matches a tool call, how Straza picks the rule that decides, and what each outcome does. - [Approvals](https://docs.straza.ai/concepts/approvals-model/index.md): Who decides a call that a rule holds, where they decide it, how long it waits, and what the agent meets in a hook and at the gateway. - [Drafts](https://docs.straza.ai/concepts/drafts/index.md): Why each change to servers, roles and policy waits as a checked draft until a person publishes it, and who may propose one. - [People, agents and sponsors](https://docs.straza.ai/concepts/identities/index.md): Who a governed action belongs to, how Straza tells a person from an AI agent, and which person is accountable for an agent. - [Machines and sessions](https://docs.straza.ai/concepts/machines-and-sessions/index.md): How a person's machine enrolls with a key that never leaves it, how each session gets a short-lived token, and how the kill switch revokes sessions and machines. - [Servers and the gateway](https://docs.straza.ai/concepts/mcp-apps-and-the-gateway/index.md): How an MCP server sits behind Straza's gateway, which tools a session sees, and what the gateway does with each tool call. - [Credentials](https://docs.straza.ai/concepts/credentials/index.md): Whose credential a call to an MCP server carries, and why the agent never holds it. - [Audit and evidence](https://docs.straza.ai/concepts/evidence/index.md): What each audit record holds, how the hash chain shows an edit or a removal, how records reach your SIEM, and what an auditor can prove with them. ## Guides - [Guides](https://docs.straza.ai/guides/index.md): Step-by-step tasks for running Straza, connecting identity, governing agents, serving MCP servers, writing policy, publishing changes, approving, using the console and forwarding the audit record. - [Set up Straza for a team](https://docs.straza.ai/guides/set-up-straza-for-a-team/index.md): Every step from an empty host to a governed agent on the enterprise profile, in the order the steps depend on each other. - [Run Straza](https://docs.straza.ai/guides/operate/index.md): Deploy, expose, back up, upgrade and debug strazad in every supported shape, and stop an identity or every agent when you must. - [Install](https://docs.straza.ai/guides/operate/install/index.md): Put the three Straza binaries on your PATH and check that they came from the Straza release job. - [Standalone host](https://docs.straza.ai/guides/operate/standalone-host/index.md): Run strazad on one Linux host in the standalone profile, with its settings in a file and all its state in one directory. - [Docker](https://docs.straza.ai/guides/operate/docker/index.md): Run a standalone strazad in a container, with all of its state on one named volume you can back up. - [The enterprise shape](https://docs.straza.ai/guides/operate/enterprise-shape/index.md): Run strazad as identical replicas behind one load balancer, all sharing one database, one event bus and one key file. - [Demo stack defaults and the hardened shape](https://docs.straza.ai/guides/operate/demo-stack-defaults/index.md): Every lab shortcut of the demo stack, the hardened shape that keeps it on the host's loopback, and what keeps it running for a team. - [Kubernetes with Helm](https://docs.straza.ai/guides/operate/kubernetes-with-helm/index.md): Install strazad on your cluster with the Helm chart, set the values that decide its shape, and know the way to upgrade and roll back. - [Run it on Kubernetes](https://docs.straza.ai/guides/operate/run-it-on-kubernetes/index.md): Run strazad on a throwaway kind cluster through the Helm chart, sign in through a port-forward, and read one deny with its reason. - [TLS and exposure](https://docs.straza.ai/guides/operate/tls-and-exposure/index.md): Serve strazad over HTTPS, give phones their own TLS listener, and keep every route that should stay private off the internet. - [Backup and upgrade](https://docs.straza.ai/guides/operate/backup-and-upgrade/index.md): Take a backup that restores the whole server, restore it and rehearse the restore, upgrade strazad without logging anyone out, and know the way back. - [Kill-switch runbook](https://docs.straza.ai/guides/operate/kill-switch-runbook/index.md): Stop one identity or stand down every session, and confirm the stop landed. - [Doctor and logs](https://docs.straza.ai/guides/operate/doctor-and-logs/index.md): Find out why a hook denies or a server misbehaves, from straza doctor on the agent's machine and the strazad log on the server. - [Metrics and alerts](https://docs.straza.ai/guides/operate/metrics-and-alerts/index.md): Scrape strazad's metrics with a token, read what each one counts, and alert on the counters that mean a person must act. - [Delegated admin](https://docs.straza.ai/guides/operate/delegated-admin/index.md): Give a person admin rights over chosen areas of Straza, such as the audit chain or sessions, without the root role. - [Delegate one MCP server](https://docs.straza.ai/guides/operate/delegate-one-server/index.md): Let the people who own one MCP server administer it and nothing else, through the server admin role Straza creates with it. - [Connect identity](https://docs.straza.ai/guides/connect-identity/index.md): Bring people, AI agents and their role membership in from your identity manager, which stays the master of who exists, while the roles themselves are created in Straza. - [midPoint](https://docs.straza.ai/guides/connect-identity/midpoint/index.md): midPoint provisions people and AI agents into Straza over SCIM, hands them Straza roles through access packages, and takes both away again. - [Okta](https://docs.straza.ai/guides/connect-identity/okta/index.md): Okta provisions people and their Straza role membership over SCIM, and you know what each request Okta sends does in Straza. - [Keycloak login](https://docs.straza.ai/guides/connect-identity/keycloak-login/index.md): strazad trusts your Keycloak realm, people sign in to Straza with their own Keycloak identity, and Straza never holds a password. - [Entra ID login](https://docs.straza.ai/guides/connect-identity/entra-login/index.md): People sign in to Straza with their Microsoft Entra ID account, in a browser on the machine they enroll, and strazad trusts your tenant's ID tokens. - [Okta login](https://docs.straza.ai/guides/connect-identity/okta-login/index.md): People sign in to Straza with their Okta account, in a browser on the machine they enroll, and strazad trusts your Okta authorization server's ID tokens. - [Generic SCIM](https://docs.straza.ai/guides/connect-identity/generic-scim/index.md): Your SCIM 2.0 client holds a token for Straza, and you know each request Straza accepts, what it does and how Straza refuses the rest. - [AI agents as identities](https://docs.straza.ai/guides/connect-identity/nhi-provisioning/index.md): An AI agent provisioned by your identity manager signs in to Straza with its own key and no browser, and leaves again when the identity manager retires it. - [Govern agents](https://docs.straza.ai/guides/govern-an-agent/index.md): Put one agent under Straza policy, harness by harness. - [Enroll a machine](https://docs.straza.ai/guides/govern-an-agent/enroll-a-machine/index.md): The Straza background service on this machine holds a key for your Straza identity, and straza status and straza doctor show it ready for an agent. - [Deploy with Intune](https://docs.straza.ai/guides/govern-an-agent/deploy-with-intune/index.md): Ship the Straza MSI to Windows machines as an Intune Win32 app that runs in System context, and the signed Straza pkg to Macs as a macOS PKG app with a shell script and a Managed Login Items profile, each with the managed install for your Straza server. - [Deploy with Jamf Pro](https://docs.straza.ai/guides/govern-an-agent/deploy-with-jamf/index.md): Ship the signed Straza pkg to Macs with one Jamf Pro policy that installs it and runs the managed install for your Straza server, and keep the service switched on with a Managed Login Items profile. - [Claude Code](https://docs.straza.ai/guides/govern-an-agent/claude-code/index.md): Claude Code on this machine hands every tool call to your Straza policy, and you have seen a destructive command denied with a reason. - [Codex CLI](https://docs.straza.ai/guides/govern-an-agent/codex-cli/index.md): Codex CLI on this machine hands every tool call to your Straza policy, and you have seen a destructive command denied with a reason. - [Gemini CLI](https://docs.straza.ai/guides/govern-an-agent/gemini-cli/index.md): Gemini CLI on this machine hands every tool call to your Straza policy, and you have seen a destructive command denied with a reason. - [Python agents](https://docs.straza.ai/guides/govern-an-agent/python-agents/index.md): Govern a Python agent with the Straza agent kit so every tool call passes through policy. - [Any MCP client](https://docs.straza.ai/guides/govern-an-agent/any-mcp-client/index.md): Your MCP client reaches its tools through the Straza gateway, and you have seen what the gateway answers when it allows a call and when it holds one for a person. - [Hookless processes](https://docs.straza.ai/guides/govern-an-agent/hookless-processes/index.md): A process with no hook surface runs inside the Straza sandbox image, where every command it runs meets your policy and its obvious bypasses fail. - [Headless agents](https://docs.straza.ai/guides/govern-an-agent/headless-agents/index.md): An AI agent with no person at the keyboard holds a key of its own, enrolls without a browser, and has a destructive command denied by your policy. - [What the agent reads back](https://docs.straza.ai/guides/govern-an-agent/what-the-agent-reads-back/index.md): The exact words an agent reads from Straza, from the session banner to a deny, a pending approval, a ticket, the built-in approval tools and whoami. - [Knowledge packs](https://docs.straza.ai/guides/govern-an-agent/knowledge-packs/index.md): A role carries a knowledge pack, and every agent session of that role starts with the pack's text in its context. - [MCP servers](https://docs.straza.ai/guides/serve-mcp-apps/index.md): Register MCP servers with Straza, hold their credentials and decide which role reaches which tool. - [Add a server](https://docs.straza.ai/guides/serve-mcp-apps/add-a-server/index.md): Register an MCP server, check that Straza reaches it, and give a role access to its tools. - [Credentials](https://docs.straza.ai/guides/serve-mcp-apps/credentials/index.md): Store one shared secret for an MCP server, so the gateway adds it to every call and no agent ever sees it. - [Each caller's own credential](https://docs.straza.ai/guides/serve-mcp-apps/caller-credentials/index.md): Each person reaches an upstream MCP server with their own token or sign-in, and you decide what agents run on. - [Access per role](https://docs.straza.ai/guides/serve-mcp-apps/catalogs-per-role/index.md): One application role reaches chosen tools of one MCP server, and you have read what its sessions see and why. - [Limit a role to some objects](https://docs.straza.ai/guides/serve-mcp-apps/limit-a-role-to-some-objects/index.md): Pin the repositories ghapp reaches, give Straza its listing credential, then give each role the repositories it covers, so an agent with write tools writes only where one of its roles allows. - [Git through Straza](https://docs.straza.ai/guides/serve-mcp-apps/git-through-straza/index.md): Let agents run plain git against GitHub with no GitHub credential on their machines, while Straza decides each fetch and push on ghapp with the roles, repositories and approvals of any tool call. - [Show MCP Apps views](https://docs.straza.ai/guides/serve-mcp-apps/show-views/index.md): Let one MCP server show its interactive views in a chat app, with every click decided, held or refused like any other tool call. - [The apps directory](https://docs.straza.ai/guides/serve-mcp-apps/gitops-apps-directory/index.md): A manifest file you keep under version control proposes a draft, a person publishes it, and deleting the file proposes the server's removal. - [Policies](https://docs.straza.ai/guides/write-policy/index.md): Write the rules that decide what an agent may do. - [Your first deny](https://docs.straza.ai/guides/write-policy/first-deny/index.md): One role's agents cannot run a recursive force-delete, and each denial tells the agent why. - [Hold a call for a person](https://docs.straza.ai/guides/write-policy/hold-for-a-human/index.md): The command kubectl apply waits for a holder of an approver role, and the agent runs it once after the approval. - [Tickets](https://docs.straza.ai/guides/write-policy/tickets/index.md): The command terraform apply needs a ticket, an approver grants it once, and a later session runs it once. - [Classify](https://docs.straza.ai/guides/write-policy/classify/index.md): Send a call to the inline classifier when a static rule cannot tell whether it is safe. - [Record a conversation](https://docs.straza.ai/guides/write-policy/capture/index.md): One role's conversations are recorded with secrets masked, and you have read a recorded prompt back. - [Simulate a call](https://docs.straza.ai/guides/write-policy/simulate-and-coverage/index.md): You have asked the live policy about one call, tested a change before it is live, replayed a recorded call, and read which policies govern each role. - [Changes and drafts](https://docs.straza.ai/guides/changes/index.md): Propose a change to MCP servers, roles, access rows and policy sets as a draft, and decide who publishes it. - [Propose and publish a change](https://docs.straza.ai/guides/changes/propose-and-publish/index.md): A change to an MCP server, a role or a policy set waits as a checked draft until a person publishes it, and a published change can be taken back with a second draft. - [Who may draft and publish](https://docs.straza.ai/guides/changes/who-may-draft/index.md): Decide who may draft a change to servers, roles, access rows and policy sets, who publishes it, and when a second person must. - [Approvals](https://docs.straza.ai/guides/approve/index.md): Answer held calls in the console, on a phone or in Slack, and make sure the request reaches the approver's phone. - [Approve in the console](https://docs.straza.ai/guides/approve/console/index.md): You approve a held call from the console, and the agent's retry runs on your decision. strazactl lists the requests and reads the decisions. - [Approve in the browser](https://docs.straza.ai/guides/approve/browser/index.md): A person who is not an admin enables their own browser on the self-service page and decides the requests routed to them, each decision signed by a key that never leaves that browser. - [Slack](https://docs.straza.ai/guides/approve/slack/index.md): strazad posts each held call to a Slack channel as a card, and a tap on Approve or Deny there decides it. - [Approve on a phone](https://docs.straza.ai/guides/approve/phone/index.md): A person enrolls their own phone as an approver device on the self-service page, decides a held call in the Straza approver app, and signs each answer with its own key. - [Push delivery and connectivity](https://docs.straza.ai/guides/approve/push-and-connectivity/index.md): Every enrolled phone can reach the approver listener, push wakes it for each request, and you have checked both from the server. - [Console](https://docs.straza.ai/guides/console/index.md): Find your way around the screens of the admin console and what each one shows. - [Overview and Audit](https://docs.straza.ai/guides/console/overview-and-audit/index.md): You read the deployment's health on Overview, then find any decision on Audit, check it against the hash chain and test the same call again. - [Roles](https://docs.straza.ai/guides/console/roles/index.md): You read what a role gives and who holds it on the role's own page, and you change its access, its composition, its holders and its knowledge packs from there. - [Policies](https://docs.straza.ai/guides/console/policies/index.md): You read which rules decide each call, build a policy with the New policy wizard, change a policy on its own page, and test a call before and after you publish. - [Settings](https://docs.straza.ai/guides/console/settings/index.md): You read every configuration value with the key that sets it, compose a config change to paste into your deployment, keep the attestation registry, mint and revoke admin API tokens, and rotate the client assertion key. - [Audit and SIEM](https://docs.straza.ai/guides/audit/index.md): Forward the audit record to a SIEM or a file and confirm that nothing is lost on the way. - [Sinks and SIEM](https://docs.straza.ai/guides/audit/sinks-and-siem/index.md): Send the audit record to a SIEM or a file sink and confirm nothing is lost on the way. - [The sentinel](https://docs.straza.ai/guides/audit/sentinel/index.md): The audit sentinel watches each session's decisions as they reach the audit chain and records a verdict when an agent probes the policy edge, and you read and act on the verdicts in Audit. ## Reference - [Reference](https://docs.straza.ai/reference/index.md): The exact surfaces, generated from the code where they can be. - [Requirements and sizing](https://docs.straza.ai/reference/requirements-and-sizing/index.md): The processors, memory and disk each shape of Straza needs, what each number rests on, how the audit record grows, and the platforms and versions Straza was tested with. - [Standalone and enterprise compared](https://docs.straza.ai/reference/standalone-and-enterprise/index.md): Every difference between the two profiles in one table, with the setting that changes each one. - [Lifetimes and timeouts](https://docs.straza.ai/reference/lifetimes-and-timeouts/index.md): How long each token, credential, session, approval and record lasts, the setting that changes it, and what happens when it runs out. - [PolicySet grammar](https://docs.straza.ai/reference/policyset-grammar/index.md): Every key of a PolicySet document, with its type, its default and a one-line example. - [Client environment](https://docs.straza.ai/reference/client-environment/index.md): Every environment variable the straza client reads, what it changes, and who sets it. - [Configuration](https://docs.straza.ai/reference/configuration/index.md): Every strazad configuration key, its environment variable when it has one, and the reason when it has none. - [API](https://docs.straza.ai/reference/api/index.md): Every operation of the strazad HTTP API by method and path, with its summary, the credential it takes, its request body, its answers and the objects they carry, from the OpenAPI document. - [Events and the audit record](https://docs.straza.ai/reference/events/index.md): List every audit event Straza emits and the fields of the audit record. - [Ports and network](https://docs.straza.ai/reference/ports-and-network/index.md): The listeners strazad opens, what each address on them serves, and every connection a deployment makes in each direction. - [The spec](https://docs.straza.ai/reference/spec/index.md): The eleven wire-format specifications Straza publishes, what each one fixes, and how to prove an implementation against them. - [Object marks for server authors](https://docs.straza.ai/reference/object-marks/index.md): The two tool marks an MCP server sends so Straza can limit a role to some of the server's objects, the listing tool's answer, the listing credential, what Straza sends on each call and what the server promises. - [Glossary](https://docs.straza.ai/reference/glossary/index.md): Each word the docs use, in one sentence, with a link to the page that explains it. ## Security - [Security](https://docs.straza.ai/security/index.md): The security model, the credential and session lifecycle, every key and token in one place, the hardening steps, how to report a vulnerability and the compliance mapping. - [The security model in depth](https://docs.straza.ai/security/security-model/index.md): What Straza defends, where each defense holds, the invariants behind it with what each one costs, and the mechanism behind the limits it leaves with you. - [Known limits](https://docs.straza.ai/security/known-limits/index.md): Every limit of what Straza protects in one table, with what it means for you, what reduces it today and what the roadmap plans, and the scale ceilings of one deployment. - [Hardening](https://docs.straza.ai/security/hardening/index.md): Every security setting checked against its default in both profiles, and each control the profiles leave to you turned on, with the command that proves it. - [Credentials and sessions](https://docs.straza.ai/security/credentials-and-sessions/index.md): Which credential sits where on a machine, how long each one lives, how a session renews, and how to stop each one. - [Keys, certificates and tokens](https://docs.straza.ai/security/keys-certificates-and-tokens/index.md): Every key and token Straza holds, where it lives, how long it lasts, how to rotate or revoke it, and what a stolen copy yields. - [Reporting a vulnerability](https://docs.straza.ai/security/reporting/index.md): Where to report a vulnerability in private, what a useful report holds, what happens after you send it, and what is in scope. - [Compliance mapping](https://docs.straza.ai/security/compliance/index.md): Which NIS2 Article 21(2) measures and ISO/IEC 27001:2022 Annex A controls Straza supplies evidence for, and the mechanism behind each one. - [Supply chain](https://docs.straza.ai/security/supply-chain/index.md): What a release holds, how to verify its signatures and the macOS pkg's Developer ID signature and notarization, how to rebuild its binaries from source, and what CI checks before a tag. ## Project - [Project](https://docs.straza.ai/project/index.md): Roadmap, changelog, contributing, support and license. - [Roadmap](https://docs.straza.ai/project/roadmap/index.md): Say what the latest Straza release added, what is confirmed for the next two months and what is planned after that. - [Changelog](https://docs.straza.ai/project/changelog/index.md): List what changed in each release. - [Contributing](https://docs.straza.ai/project/contributing/index.md): Tell you how to propose a change and what a contribution needs before it lands. - [Support](https://docs.straza.ai/project/support/index.md): Tell you where to ask for help and what to include when you do. - [FAQ](https://docs.straza.ai/project/faq/index.md): Short answers to the questions people ask first about Straza, each with a link to the page that covers its topic. - [License](https://docs.straza.ai/project/license/index.md): State the license terms Straza ships under. ## Optional - [Command line](https://docs.straza.ai/reference/cli/index.md): Every command of strazad, strazactl and straza with its flags and operands. - [strazad](https://docs.straza.ai/reference/cli/strazad/index.md): Straza platform server - [strazad migrate](https://docs.straza.ai/reference/cli/strazad/strazad_migrate/index.md): Move the database schema to one migration version - [strazad serve](https://docs.straza.ai/reference/cli/strazad/strazad_serve/index.md): Run the platform server - [strazad version](https://docs.straza.ai/reference/cli/strazad/strazad_version/index.md): Print version information - [strazactl](https://docs.straza.ai/reference/cli/strazactl/index.md): Straza admin CLI - [strazactl api-token](https://docs.straza.ai/reference/cli/strazactl/strazactl_api-token/index.md): Manage long-lived admin API tokens, the one credential for automation, connectors and an identity manager's SCIM push - [strazactl api-token create](https://docs.straza.ai/reference/cli/strazactl/strazactl_api-token_create/index.md): Mint a long-lived admin API token (shown exactly once) - [strazactl api-token list](https://docs.straza.ai/reference/cli/strazactl/strazactl_api-token_list/index.md): List admin API token metadata - [strazactl api-token revoke](https://docs.straza.ai/reference/cli/strazactl/strazactl_api-token_revoke/index.md): Revoke an admin API token by its name or its id - [strazactl approvals](https://docs.straza.ai/reference/cli/strazactl/strazactl_approvals/index.md): List approval requests - [strazactl approvals list](https://docs.straza.ai/reference/cli/strazactl/strazactl_approvals_list/index.md): List approval requests (default: pending) - [strazactl approvers](https://docs.straza.ai/reference/cli/strazactl/strazactl_approvers/index.md): Inspect and revoke enrolled approver phones and browsers - [strazactl approvers list](https://docs.straza.ai/reference/cli/strazactl/strazactl_approvers_list/index.md): List enrolled approver phones and browsers, with owner, key posture and push routes - [strazactl approvers revoke](https://docs.straza.ai/reference/cli/strazactl/strazactl_approvers_revoke/index.md): Revoke an enrolled approver device (its approver credential fails on the next call) - [strazactl apps](https://docs.straza.ai/reference/cli/strazactl/strazactl_apps/index.md): Manage MCP servers - [strazactl apps bind](https://docs.straza.ai/reference/cli/strazactl/strazactl_apps_bind/index.md): Give a role the server owns access to its tools (a tool without access does not exist for the role) - [strazactl apps disable](https://docs.straza.ai/reference/cli/strazactl/strazactl_apps_disable/index.md): Pause a server: stop it and keep it stopped until re-enabled - [strazactl apps enable](https://docs.straza.ai/reference/cli/strazactl/strazactl_apps_enable/index.md): Clear a server's admin pause and (re)start it - [strazactl apps export](https://docs.straza.ai/reference/cli/strazactl/strazactl_apps_export/index.md): Print a server's stored manifest as app.yaml on stdout, the document apps install takes - [strazactl apps import](https://docs.straza.ai/reference/cli/strazactl/strazactl_apps_import/index.md): Convert an MCP registry server.json into an App manifest, by the registry import rules of the app manifest specification - [strazactl apps install](https://docs.straza.ai/reference/cli/strazactl/strazactl_apps_install/index.md): Install or update MCP servers from app.yaml manifests - [strazactl apps list](https://docs.straza.ai/reference/cli/strazactl/strazactl_apps_list/index.md): List MCP servers with live status, the reason behind it and the roles that reach each - [strazactl apps logs](https://docs.straza.ai/reference/cli/strazactl/strazactl_apps_logs/index.md): Show recent runtime log lines for a server - [strazactl apps objects](https://docs.straza.ai/reference/cli/strazactl/strazactl_apps_objects/index.md): Show what Straza last listed on a server, such as its repos, and the roles that cover each - [strazactl apps objects draft](https://docs.straza.ai/reference/cli/strazactl/strazactl_apps_objects_draft/index.md): Open a draft that pins the objects a server declares in its tool marks - [strazactl apps objects refresh](https://docs.straza.ai/reference/cli/strazactl/strazactl_apps_objects_refresh/index.md): List a server's objects again now instead of at the next 10-minute listing - [strazactl apps recheck](https://docs.straza.ai/reference/cli/strazactl/strazactl_apps_recheck/index.md): Trigger an immediate health probe and print the refreshed state - [strazactl apps remove](https://docs.straza.ai/reference/cli/strazactl/strazactl_apps_remove/index.md): Stop and remove a server - [strazactl apps secret](https://docs.straza.ai/reference/cli/strazactl/strazactl_apps_secret/index.md): Manage a server's static secret (injected gateway-side, never client-visible) - [strazactl apps secret remove](https://docs.straza.ai/reference/cli/strazactl/strazactl_apps_secret_remove/index.md): Remove the server's own secret, with --role the override one role uses, or with --listing the listing credential - [strazactl apps secret set](https://docs.straza.ai/reference/cli/strazactl/strazactl_apps_secret_set/index.md): Store the server's own secret, with --role the override one role uses instead, or with --listing the listing credential - [strazactl apps show](https://docs.straza.ai/reference/cli/strazactl/strazactl_apps_show/index.md): Show one server in full: health and its reason, tools, who has access and how each tool runs, the stored secrets - [strazactl apps tools](https://docs.straza.ai/reference/cli/strazactl/strazactl_apps_tools/index.md): List exposed MCP tools with their upstream descriptions, every server or one - [strazactl apps unbind](https://docs.straza.ai/reference/cli/strazactl/strazactl_apps_unbind/index.md): Remove a role's access to a server (the id is in bindings list and apps show) - [strazactl assign](https://docs.straza.ai/reference/cli/strazactl/strazactl_assign/index.md): Assign a role to a user - [strazactl attestation](https://docs.straza.ai/reference/cli/strazactl/strazactl_attestation/index.md): Manage the managed-install expected-hash registry - [strazactl attestation add](https://docs.straza.ai/reference/cli/strazactl/strazactl_attestation_add/index.md): Register an expected measurement (allowed-set entry) - [strazactl attestation list](https://docs.straza.ai/reference/cli/strazactl/strazactl_attestation_list/index.md): List registered measurements - [strazactl attestation rm](https://docs.straza.ai/reference/cli/strazactl/strazactl_attestation_rm/index.md): Remove a measurement from the registry - [strazactl audit](https://docs.straza.ai/reference/cli/strazactl/strazactl_audit/index.md): Query and verify the audit log - [strazactl audit tail](https://docs.straza.ai/reference/cli/strazactl/strazactl_audit_tail/index.md): Show the most recent audit records - [strazactl audit verify](https://docs.straza.ai/reference/cli/strazactl/strazactl_audit_verify/index.md): Verify the audit hash chain end to end - [strazactl bindings](https://docs.straza.ai/reference/cli/strazactl/strazactl_bindings/index.md): List the access rows of roles on servers - [strazactl bindings list](https://docs.straza.ai/reference/cli/strazactl/strazactl_bindings_list/index.md): List the access rows - [strazactl catalog](https://docs.straza.ai/reference/cli/strazactl/strazactl_catalog/index.md): Inspect the effective tool catalog - [strazactl catalog preview](https://docs.straza.ai/reference/cli/strazactl/strazactl_catalog_preview/index.md): Preview the tool catalog a subject would see: named roles, a real user's roles, or both - [strazactl connect](https://docs.straza.ai/reference/cli/strazactl/strazactl_connect/index.md): Connect a user's token to an MCP server; no server lists the user's connections - [strazactl devices](https://docs.straza.ai/reference/cli/strazactl/strazactl_devices/index.md): Inspect and revoke a user's enrolled devices - [strazactl devices list](https://docs.straza.ai/reference/cli/strazactl/strazactl_devices_list/index.md): List a user's enrolled devices - [strazactl devices revoke](https://docs.straza.ai/reference/cli/strazactl/strazactl_devices_revoke/index.md): Revoke a device enrollment (it stops starting sessions) - [strazactl disconnect](https://docs.straza.ai/reference/cli/strazactl/strazactl_disconnect/index.md): Remove a user's connection to a server - [strazactl drafts](https://docs.straza.ai/reference/cli/strazactl/strazactl_drafts/index.md): Check, store and review drafts of changes to MCP servers, roles, access rows and policy sets - [strazactl drafts check](https://docs.straza.ai/reference/cli/strazactl/strazactl_drafts_check/index.md): Check documents against live state and print the verdict, storing nothing - [strazactl drafts create](https://docs.straza.ai/reference/cli/strazactl/strazactl_drafts_create/index.md): Store the documents as a new draft and print its verdict - [strazactl drafts discard](https://docs.straza.ai/reference/cli/strazactl/strazactl_drafts_discard/index.md): Discard an open draft, which changes nothing live - [strazactl drafts list](https://docs.straza.ai/reference/cli/strazactl/strazactl_drafts_list/index.md): List drafts, newest first, the open ones unless --state names others - [strazactl drafts rebase](https://docs.straza.ai/reference/cli/strazactl/strazactl_drafts_rebase/index.md): Check a draft again on live state, keeping the fields it changes and taking every other field from live - [strazactl drafts revert](https://docs.straza.ai/reference/cli/strazactl/strazactl_drafts_revert/index.md): Make a new draft that undoes a published one, checked like any other - [strazactl drafts show](https://docs.straza.ai/reference/cli/strazactl/strazactl_drafts_show/index.md): Show a draft: who proposed it, each object against live state, its verdict and who gains what - [strazactl drafts update](https://docs.straza.ai/reference/cli/strazactl/strazactl_drafts_update/index.md): Replace an open draft's documents with a new revision and print its verdict - [strazactl packs](https://docs.straza.ai/reference/cli/strazactl/strazactl_packs/index.md): Manage knowledge packs - [strazactl packs bind](https://docs.straza.ai/reference/cli/strazactl/strazactl_packs_bind/index.md): Bind a knowledge pack to a role - [strazactl packs create](https://docs.straza.ai/reference/cli/strazactl/strazactl_packs_create/index.md): Create a knowledge pack from --content or --file - [strazactl packs delete](https://docs.straza.ai/reference/cli/strazactl/strazactl_packs_delete/index.md): Delete a knowledge pack that no role is bound to (unbind it first) - [strazactl packs list](https://docs.straza.ai/reference/cli/strazactl/strazactl_packs_list/index.md): List knowledge packs - [strazactl packs unbind](https://docs.straza.ai/reference/cli/strazactl/strazactl_packs_unbind/index.md): Unbind a knowledge pack from a role - [strazactl policy](https://docs.straza.ai/reference/cli/strazactl/strazactl_policy/index.md): Validate and manage PolicySets - [strazactl policy activate](https://docs.straza.ai/reference/cli/strazactl/strazactl_policy_activate/index.md): Publish one PolicySet's stored text after the server's checks, leaving every other set as published - [strazactl policy apply](https://docs.straza.ai/reference/cli/strazactl/strazactl_policy_apply/index.md): Upload PolicySet YAML: a new set is stored off, and edits to a live set wait for policy activate - [strazactl policy deactivate](https://docs.straza.ai/reference/cli/strazactl/strazactl_policy_deactivate/index.md): Turn a PolicySet off and recompile and distribute the snapshot - [strazactl policy delete](https://docs.straza.ai/reference/cli/strazactl/strazactl_policy_delete/index.md): Delete a PolicySet that is off (turn a live one off first) - [strazactl policy diff](https://docs.straza.ai/reference/cli/strazactl/strazactl_policy_diff/index.md): Diff the stored PolicySet against a local file - [strazactl policy list](https://docs.straza.ai/reference/cli/strazactl/strazactl_policy_list/index.md): List stored PolicySets - [strazactl policy show](https://docs.straza.ai/reference/cli/strazactl/strazactl_policy_show/index.md): Print a stored PolicySet: its story on stderr, the YAML alone on stdout - [strazactl policy simulate](https://docs.straza.ai/reference/cli/strazactl/strazactl_policy_simulate/index.md): Ask what a call would do for a subject, against the live policies - [strazactl policy validate](https://docs.straza.ai/reference/cli/strazactl/strazactl_policy_validate/index.md): Validate PolicySet YAML files locally against the v1beta1 spec - [strazactl roles](https://docs.straza.ai/reference/cli/strazactl/strazactl_roles/index.md): Manage roles - [strazactl roles create](https://docs.straza.ai/reference/cli/strazactl/strazactl_roles_create/index.md): Create a role - [strazactl roles delete](https://docs.straza.ai/reference/cli/strazactl/strazactl_roles_delete/index.md): Delete a role; its assignments and access rows go with it - [strazactl roles export](https://docs.straza.ai/reference/cli/strazactl/strazactl_roles_export/index.md): Print the role's canonical VCS document (spec/objects kind Role) to stdout - [strazactl roles implications](https://docs.straza.ai/reference/cli/strazactl/strazactl_roles_implications/index.md): List the roles this role composes (holding it also holds these) - [strazactl roles implications add](https://docs.straza.ai/reference/cli/strazactl/strazactl_roles_implications_add/index.md): Compose one role into another: holding the first also holds the second - [strazactl roles implications remove](https://docs.straza.ai/reference/cli/strazactl/strazactl_roles_implications_remove/index.md): Stop composing one role into another (holding the first no longer holds the second) - [strazactl roles list](https://docs.straza.ai/reference/cli/strazactl/strazactl_roles_list/index.md): List roles - [strazactl roles update](https://docs.straza.ai/reference/cli/strazactl/strazactl_roles_update/index.md): Update a role's description (the name and the kind are fixed at create) - [strazactl sessions](https://docs.straza.ai/reference/cli/strazactl/strazactl_sessions/index.md): Inspect and revoke sessions - [strazactl sessions list](https://docs.straza.ai/reference/cli/strazactl/strazactl_sessions_list/index.md): List sessions - [strazactl sessions revoke](https://docs.straza.ai/reference/cli/strazactl/strazactl_sessions_revoke/index.md): Revoke a session (kill switch) - [strazactl sessions transcript](https://docs.straza.ai/reference/cli/strazactl/strazactl_sessions_transcript/index.md): Print a session's recorded conversation (recording is policy-opted-in) - [strazactl signing-keys](https://docs.straza.ai/reference/cli/strazactl/strazactl_signing-keys/index.md): List, rotate, promote and retire signing keys, and print a client assertion key's certificate - [strazactl signing-keys certificate](https://docs.straza.ai/reference/cli/strazactl/strazactl_signing-keys_certificate/index.md): Print the certificate of a client assertion key, to upload to an Entra agent blueprint - [strazactl signing-keys list](https://docs.straza.ai/reference/cli/strazactl/strazactl_signing-keys_list/index.md): List every signing key with its purpose and status - [strazactl signing-keys promote](https://docs.straza.ai/reference/cli/strazactl/strazactl_signing-keys_promote/index.md): Make a staged client assertion key the signing key - [strazactl signing-keys retire](https://docs.straza.ai/reference/cli/strazactl/strazactl_signing-keys_retire/index.md): Retire one client assertion key at once, for a key that may have been copied - [strazactl signing-keys rotate](https://docs.straza.ai/reference/cli/strazactl/strazactl_signing-keys_rotate/index.md): Stage a new key of one purpose: a session key promotes itself, a client assertion key waits for promote - [strazactl sinks](https://docs.straza.ai/reference/cli/strazactl/strazactl_sinks/index.md): Inspect SIEM sinks and replay parked (dead-letter) events - [strazactl sinks list](https://docs.straza.ai/reference/cli/strazactl/strazactl_sinks_list/index.md): List configured sinks with backlog, parked count, and the last failure - [strazactl sinks replay](https://docs.straza.ai/reference/cli/strazactl/strazactl_sinks_replay/index.md): Re-deliver a sink's parked events to its receiver, oldest first - [strazactl spec](https://docs.straza.ai/reference/cli/strazactl/strazactl_spec/index.md): Work with the Straza spec artifacts - [strazactl spec conformance](https://docs.straza.ai/reference/cli/strazactl/strazactl_spec_conformance/index.md): Replay the published Tier-1 corpus against a hook implementation and score it - [strazactl spec validate](https://docs.straza.ai/reference/cli/strazactl/strazactl_spec_validate/index.md): Validate documents against the embedded v1beta1 spec validators - [strazactl status](https://docs.straza.ai/reference/cli/strazactl/strazactl_status/index.md): Report the resolved target server, then strazad health and version - [strazactl transcripts](https://docs.straza.ai/reference/cli/strazactl/strazactl_transcripts/index.md): Search recorded conversations by text or by a secret's value - [strazactl transcripts search](https://docs.straza.ai/reference/cli/strazactl/strazactl_transcripts_search/index.md): Find who said what: substring scan and/or exact-value match - [strazactl unassign](https://docs.straza.ai/reference/cli/strazactl/strazactl_unassign/index.md): Take a role away from a user - [strazactl users](https://docs.straza.ai/reference/cli/strazactl/strazactl_users/index.md): Manage users - [strazactl users create](https://docs.straza.ai/reference/cli/strazactl/strazactl_users_create/index.md): Create a local user - [strazactl users create-nhi](https://docs.straza.ai/reference/cli/strazactl/strazactl_users_create-nhi/index.md): Provision an AI agent or a service account (headless, no password, key-based) - [strazactl users disable](https://docs.straza.ai/reference/cli/strazactl/strazactl_users_disable/index.md): Disable a user (status disabled, every session revoked, lifted by `strazactl users enable`) - [strazactl users enable](https://docs.straza.ai/reference/cli/strazactl/strazactl_users_enable/index.md): Enable a disabled user: the status goes back to active, the user's revocations are lifted, and a device still enrolled needs no new enrollment - [strazactl users entra-agent-id](https://docs.straza.ai/reference/cli/strazactl/strazactl_users_entra-agent-id/index.md): Manage the Entra agent identity an AI agent acts as at Entra - [strazactl users entra-agent-id set](https://docs.straza.ai/reference/cli/strazactl/strazactl_users_entra-agent-id_set/index.md): Set the client id of the Entra agent identity the agent acts as - [strazactl users entra-agent-id show](https://docs.straza.ai/reference/cli/strazactl/strazactl_users_entra-agent-id_show/index.md): Show the Entra agent identity the agent acts as - [strazactl users entra-agent-id unset](https://docs.straza.ai/reference/cli/strazactl/strazactl_users_entra-agent-id_unset/index.md): Clear the agent's Entra agent identity - [strazactl users list](https://docs.straza.ai/reference/cli/strazactl/strazactl_users_list/index.md): List users - [strazactl users lock](https://docs.straza.ai/reference/cli/strazactl/strazactl_users_lock/index.md): Lock a user: a revocation that an identity manager cannot lift by enabling the user; the status stays as it is - [strazactl users nhi-key](https://docs.straza.ai/reference/cli/strazactl/strazactl_users_nhi-key/index.md): Manage an agent identity's headless assertion key (straza keygen prints it) - [strazactl users nhi-key set](https://docs.straza.ai/reference/cli/strazactl/strazactl_users_nhi-key_set/index.md): Register the agent identity's Ed25519 public key, or rotate it and end the agent's running sessions - [strazactl users nhi-key unset](https://docs.straza.ai/reference/cli/strazactl/strazactl_users_nhi-key_unset/index.md): Revoke the agent identity's key, which ends its running sessions and refuses its next token request - [strazactl users set-password](https://docs.straza.ai/reference/cli/strazactl/strazactl_users_set-password/index.md): Set a person's built-in-issuer password (rotation, or a credential for SCIM-born people) - [strazactl users unlock](https://docs.straza.ai/reference/cli/strazactl/strazactl_users_unlock/index.md): Unlock a user: lift the user's revocations and leave the status as it is, so a disabled user still needs `strazactl users enable` - [strazactl version](https://docs.straza.ai/reference/cli/strazactl/strazactl_version/index.md): Print version information - [straza](https://docs.straza.ai/reference/cli/straza/index.md): Straza client: enroll this machine, wire the hooks and decide each tool call - [straza connect](https://docs.straza.ai/reference/cli/straza/straza_connect/index.md): Connect your own account to an MCP server; no server lists your connections - [straza disconnect](https://docs.straza.ai/reference/cli/straza/straza_disconnect/index.md): Remove your own connection to an MCP server - [straza doctor](https://docs.straza.ai/reference/cli/straza/straza_doctor/index.md): Diagnose the background service, enrollment, connectivity, hook wiring, and the audit spool - [straza enroll](https://docs.straza.ai/reference/cli/straza/straza_enroll/index.md): Sign in through your browser and enroll this machine with the background service (--use-device-code: sign in with a code, --headless: an AI agent) - [straza exec](https://docs.straza.ai/reference/cli/straza/straza_exec/index.md): Run a command through Straza policy, for an agent that has no hooks - [straza hook](https://docs.straza.ai/reference/cli/straza/straza_hook/index.md): Hook entrypoint: normalize a harness payload and decide (reads stdin) - [straza install](https://docs.straza.ai/reference/cli/straza/straza_install/index.md): Write hook wiring for Tier-1 harnesses (user mode, or --managed system layout) - [straza keygen](https://docs.straza.ai/reference/cli/straza/straza_keygen/index.md): Have the background service generate the Ed25519 key an AI agent enrolls with headless - [straza logs](https://docs.straza.ai/reference/cli/straza/straza_logs/index.md): Print the client error log (hook, capture and MCP bridge failures; never payload content) - [straza mcp](https://docs.straza.ai/reference/cli/straza/straza_mcp/index.md): Serve the Straza MCP gateway to this harness over stdio (`straza install` registers it) - [straza service](https://docs.straza.ai/reference/cli/straza/straza_service/index.md): Run the Straza background service that holds this machine's keys and sessions - [straza status](https://docs.straza.ai/reference/cli/straza/straza_status/index.md): Show enrollment, sessions and the audit spool the background service holds for you - [straza trace](https://docs.straza.ai/reference/cli/straza/straza_trace/index.md): Decision journal and debug trace (content-free): on, off, status, show - [straza trace off](https://docs.straza.ai/reference/cli/straza/straza_trace_off/index.md): Close the debug window (the journal stays on) - [straza trace on](https://docs.straza.ai/reference/cli/straza/straza_trace_on/index.md): Open a debug window (default 1h, max 24h); the journal is always on - [straza trace show](https://docs.straza.ai/reference/cli/straza/straza_trace_show/index.md): Print the journal and debug records, oldest first - [straza trace status](https://docs.straza.ai/reference/cli/straza/straza_trace_status/index.md): Report the trace level in force, its source, the window and the file - [straza uninstall](https://docs.straza.ai/reference/cli/straza/straza_uninstall/index.md): Remove Straza hook wiring from harness settings - [straza version](https://docs.straza.ai/reference/cli/straza/straza_version/index.md): Print version information