Bragproof

← Reports

@syntica/mcp

Starts with no setup2.4.1npm

Installed, started with no arguments, no environment and no input, and listed its tools. Run 2026-11-16.

Needs to start: nothing
Returns a result when called with no setup: 1 of 4 tools tried
Contacts while installing and starting: nothing

82tools listed
0hosts contacted
0files written
0.2 sto start
95dependencies

What it contacted

Nothing besides the npm registry it was installed from. Watched from outside its sandbox while it was installed and started with no input. Its tools may still contact things when called.

Starting it

npx -y @syntica/mcp@2.4.1

Installing a package runs its code. Use a sandbox for anything you have not read.

Calling its tools

4 tools that take no input and look read-only were called once with nothing. 1 returned a result.

org_list_orgsNo usable answer tool reported an error: {"success":false,"data":null,"error":{"code":"NOT_SIGNED_IN","message":"Not signed in to Syntica on this machine.","hint":"Ask the user
org_list_plansReturned a result 0.5 s · 39 characters
validation_get_generic_promptsNo usable answer tool reported an error: {"success":false,"data":null,"error":{"code":"NOT_SIGNED_IN","message":"Not signed in to Syntica on this machine.","hint":"Ask the user
validation_list_promptsNo usable answer tool reported an error: {"success":false,"data":null,"error":{"code":"NOT_SIGNED_IN","message":"Not signed in to Syntica on this machine.","hint":"Ask the user

During the calls (not necessarily because of them) it first contacted: app.syntica.io.

Only tools whose names read like looking something up, with no required input, are tried, five at most. Tools that need input, and anything that sounds like it changes or sends something, are not called.

What its tools tell an agent

In the tool descriptions, nothing that talks about the user, other tools or private files, or uses hidden characters.

1 tool with commands in capitals or very long text
  • Gives the agent orders in app_describe_connections: latform injects at deploy time. Never returns a credential. ALWAYS call this before writing or changing code that talks to an external system: write against the injec

Common and usually harmless: it is how many makers stress how a tool should be used.

Found by simple pattern rules in each tool's description (up to 3,000 characters), quoted as written. Parameter descriptions are not scanned. A clean result is not a safety review.

Tools

org_authenticate
Check whether the user is signed in, and as whom. Returns the authenticated user and organisation (use its orgId for tools that take one). Call this first in a session; if it reports the user is not signed in, run the sign-in flow: org_request_otp, then org_verify_otp.
org_logout
Sign the user out of Syntica on this machine. Only when the user asks to sign out or to sign in as someone else. Every Claude session on this machine shares the login, so they are all signed out.
org_refresh_token
Refresh an existing session token, extending its validity for a full new lifetime (7 days by default). Rarely needed: tokens past half of their lifetime are refreshed automatically on any tool call. Use this only to force a refresh explicitly.
org_register name, email, orgName
Register a new organisation and its first admin user. The default subscription plan is applied automatically.
org_invite_member orgId, email, role
Invite a user to the organisation by email with a specified role
77 more tools
org_accept_invite token, name
Accept an organisation invite using the invite token received by email
org_request_otp email
Send a one-time password to the user's email address to start the login flow. Call org_verify_otp next with the code from the email.
org_verify_otp email, code
Verify the one-time password and sign the user in. The MCP server keeps the session itself: there is no token to store or pass to other tools. Returns the signed-in user and organisation. If the response has pendingOrgSelection=true, the user belongs to multiple organisations: ask them which one to use and call org_select with the chosen orgId.
org_select orgId
Choose which organisation to use after login (when org_verify_otp returned pendingOrgSelection), or switch to a different organisation mid-session. The session moves to that organisation; every later call works in it.
org_list_orgs
List the organisations the current user belongs to, with the one the current token is scoped to. Call this whenever the user mentions another organisation, asks what they have access to, or before creating an application when you have not confirmed which organisation they mean — a session token is scoped to ONE organisation, and working in another needs org_select, not a different orgId argument.
app_list orgId
List all applications for an organisation.
app_create name, orgId, templateId, githubConnectionId, githubRepoName, githubPrivate, cloudAccount
Create a new application. Returns the application ID and git clone URL for pushing code. The application is created in the organisation the token is scoped to — if the user belongs to more than one (org_list_orgs), ask which one they mean before creating, and call org_select first if it is not the current one.
app_delete applicationId, when
Delete an application together with every environment and instance it runs, and its data. The repository is moved to trash and can be restored with app_restore for about 30 days, but the deployed environments and their data are destroyed and cannot be restored. By default the delete happens now. Slots already paid for stay paid until the 1st and can be reused by a new application or environment until then, and the reply lists them. Pass when "period_end" to delete on the 1st instead, or an ISO-8601 time before the 1st. Needs an org admin. Destructive: confirm with the user before calling.
app_list_deleted orgId
List soft-deleted applications for an organisation (the restore view). Use this to find the applicationId to pass to app_restore, and to see how long is left before each is permanently purged.
app_restore applicationId
Restore a soft-deleted application within its recovery window. Find the applicationId with app_list_deleted first. Fails if the recovery window has elapsed (the app was already permanently purged).
app_rename applicationId, name
Rename an application and its git repository. WARNING: this changes the repo path, so existing git clone/remote URLs (and the cloneUrl) stop working — anyone with a local clone must update their remote to the new cloneUrl returned here. The name must be lowercase letters, numbers and hyphens only, and unique in the org.
org_list_members orgId
List all members of an organisation and their roles
org_list_plans
List the available subscription plans (free and pay-as-you-go) from the Stripe catalog, with each plan's name and base price
org_get_subscription orgId
Get the current subscription plan (free or pay-as-you-go) for an organisation
org_get_billing orgId
Show the organisation's current bill: the plan, the monthly total Stripe will charge, and a per-line breakdown (node tiers, applications, environments, backups, traffic, add-ons) with quantities and amounts. Read-only — this does not change anything.
org_setup_git_credentials label
Generate a long-lived git credential token for cloning and pushing repositories via HTTPS. Stores the token so git works from any terminal without re-authenticating. If an application deploys from your own repository (see app_set_upstream), these credentials are read-only for it — pushes to the platform-hosted repository are refused, and you should push to your own repository instead. NOT for connecting your own GitHub/GitLab/Bitbucket repository — to store a credential for a host you own repositories on, use org_add_git_connection instead.
org_revoke_git_credentials credentialId
Revoke a git credential token by its ID
org_list_git_credentials
List all git credentials for the authenticated user
app_set_upstream applicationId, connectionId, path, branch
Point an application at a git repository you host yourself, using a remote repository connection you have already created. Create that connection first with org_add_git_connection, and find its id with org_list_git_connections — the credential lives on the connection, so this call takes only the repository path within its host and the branch. WARNING: this is destructive to the platform-hosted repository. On the next deploy the platform force-fetches your repository over its mirror, so any commit that exists only in the platform-hosted repository is lost. Afterwards pushes to it are refused and templates can no longer be applied — push to your own repository instead. Returns the list of branches found in the repository.
app_clear_upstream applicationId
Disconnect an application from its own git repository and return it to a platform-hosted repository. The hosted repository becomes writable again (its history is kept). The stored access token remains in your organisation and can be removed with org_list_git_connections.
org_list_git_connections orgId
List the git connections (stored access tokens for your own repositories) in an organisation. Credential material is never returned — only the provider, host, username, last-used/last-error status, and isDefault for the org's default GitHub account (change it with org_set_default_git_connection).
org_set_default_git_connection orgId, connectionId
Set or clear the organisation's default GitHub account (org admins only). New applications created from Claude in the browser or the web app land in the default account, so members without a GitHub account can still create and edit apps. Pass a connectionId from org_list_git_connections to make it the default: it must be a Syntica GitHub App installation (appInstallation: true) on a GitHub Organization, not a personal account. Omit connectionId to clear the default. At most one connection is the default; setting a new one replaces the old.
org_add_git_connection orgId, url, username, accessToken, label, provider
Add a remote repository: store a credential for a git host you own repositories on (GitHub, GitLab, Bitbucket, or any HTTPS git server), so applications can deploy from it. This is the paid external-source-repository feature. For GitHub, prefer org_connect_github (the Syntica GitHub App): one click, no token to manage, and included on every plan. NOT for the platform-hosted repository — for a token to clone or push that, use org_setup_git_credentials instead. One connection covers every repository on that host for that username, and applications are bound to it with app_set_upstream. Re-running this for the same host and username rotates the stored token rather than creating a duplicate. Credential material is never returned.
org_connect_github orgId
Connect a GitHub account to the organisation via the Syntica GitHub App — no personal access token needed, and included on every plan. Returns an install URL the USER must open in their browser to approve the installation; once approved, the connection appears in org_list_git_connections with appInstallation: true. With such a connection, app_create can create the source repository in their GitHub account (githubConnectionId) so they can edit code on GitHub and deploy from it, and org_list_github_repos lists their existing repositories to deploy from instead. For a non-GitHub host, or to connect a pre-existing repository with a token, use org_add_git_connection instead (paid feature).
org_list_github_repos orgId, connectionId
List the repositories a GitHub App connection can reach (newest-pushed first, up to 100), so the user picks one instead of typing owner/repo. Use with a connection from org_list_git_connections that has appInstallation: true. To deploy from one of the listed repositories, call app_set_upstream with its fullName as the path and (usually) its defaultBranch — note that binding a pre-existing repository is the paid external-source-repository feature. Creating a NEW repository instead (app_create with githubConnectionId) is included on every plan. If a repository is missing here, the user limited the installation to selected repositories — they manage that on GitHub under Settings → Applications.
app_repo_tree applicationId
List the file tree of an application's code, on its deploy branch. No GitHub credential is needed in this session. Call this before reading or saving, to see what exists. Works for every app, Syntica-hosted or GitHub-source. All access is confined to the repository of this application, on its deploy branch.
app_repo_read_file applicationId, path
Read one file of an application's code, on its deploy branch. Read a file before editing it, because a save replaces whole files. Works for every app, Syntica-hosted or GitHub-source. All access is confined to the repository of this application, on its deploy branch.
app_repo_commit applicationId, message, files, deletes
Save file changes into an application's code. All changes land together as one new version on the deploy branch. Each file REPLACES the file at its path in full (read it first to edit part of it), and deletes remove paths. On a Syntica-hosted app, saving updates the dev app automatically: the result carries a deployId, and you check progress with org_deploy_status. On a GitHub-source app the save goes to GitHub, and you run org_deploy afterwards to ship it. Saving never touches staging or production, which need org_deploy or org_promote. Repeat the userMessage of the result to the user. Tell the user their change is "saved" and that you are "updating your app", and call earlier states "versions", never say "commit". Works for every app, Syntica-hosted or GitHub-source. All access is confined to the repository of this application, on its deploy branch.
app_list_versions applicationId, limit
List the saved versions of a Syntica-hosted application, newest first, with when each was saved, by whom and what changed. Use it when the user wants to go back to an earlier state, then pass a version id to app_restore_version. A GitHub-source app keeps its versions on GitHub, and this tool says so. Tell the user their change is "saved" and that you are "updating your app", and call earlier states "versions", never say "commit".
app_restore_version applicationId, versionId
Put back the files of an earlier version of a Syntica-hosted application, as a new save. Nothing is lost: the version it replaces stays in the list. Pass a versionId from app_list_versions, or leave it out to undo the last change. Like any save, it updates the dev app automatically: check progress with org_deploy_status and the deployId. Repeat the userMessage of the result to the user. Tell the user their change is "saved" and that you are "updating your app", and call earlier states "versions", never say "commit".
org_get_action_policy applicationId, environment
Check what the organisation allows an AGENT to do in one application environment before proposing destructive work. Returns, per category (data_loss, availability, credentials), whether agents may act freely ("allow"), only with an administrator's token ("admin_only"), or not at all ("block"), and which scope that came from. Deleting production is admin-only by default. Read-only: the policy itself can only be changed by an administrator in the dashboard.
app_list_node_tiers orgId, applicationId, environment, cloudAccount
List the node tiers (instance sizes) available to an organisation, so the user can choose what an application's environment should run on. Returns the org's plan key alongside the tiers. On the Pay-as-you-go plan every tier is billed and the free tier is NOT selectable — always ask the user which tier to use rather than assuming one. On the free plan the free tier is the only option. Pass applicationId and environment (or cloudAccount for a new application) to get the sizes for that environment: one in the organisation's own cloud account runs on its own AWS bill, so every size is offered whatever the plan, with inCustomerAccount true and mustAsk true. Set the chosen tier with app_set_node_tier.
app_set_node_tier applicationId, environment, nodeTierId, when, cloudAccount, regionId
Set which node tier (instance size) an application's environment runs on. Required before the first deploy of any environment on the Pay-as-you-go plan — org_deploy fails with NODE_TIER_REQUIRED until it is set, and must not be guessed: the tier determines what the customer is billed. Ask the user to choose from app_list_node_tiers first. On an already-provisioned environment the instance must be stopped first (app_stop_instance) — the change is then carried out on the infrastructure by app_reprovision_instance, after which the instance can be started again. Pass when "period_end" to schedule a downgrade for the 1st (or an ISO-8601 time before it) instead of now. The instance keeps its current size until then and nothing needs stopping. Only a downgrade can be scheduled. The reply lists paid slots that nothing uses, which can be reused until the 1st.
org_list_cloud_accounts orgId
List the cloud accounts an organisation can run environments in: the Syntica hosted account plus any AWS accounts the organisation connected. Returns id, name, provider, region, status, whether it is the default, its namespace (the name its environments get their default address under), updateRequired and placeable. Only an account with placeable true takes new environments. An own account becomes placeable once it is connected and its stack is recent enough. When a newer stack template is available, an account also carries updateUrl, the link that opens the stack update in the AWS console with the new template filled in, and whatChanges, one plain sentence per template version with what it unlocks. Updating is optional unless the organisation wants what it unlocks. A person applies it in the AWS console (review, tick the IAM acknowledgement, Submit). Syntica never updates the stack and this server cannot. An account that is still connecting (status pending) carries connection: the step it is at (linkOpened, stackStarted, networkCreated, roleCreated, verifying), its state, when it started, the stack link once known, and, when it failed or its link expired, why. A failed or expired connection is started again by an admin in the web app. Pass an id or name as cloudAccount to app_create, app_enable_environment or app_set_node_tier. Connecting or removing an account is done by an admin in the web app (Settings, Cloud accounts), never through this server. Needs cloud accounts to be enabled for the organisation. Without that, leave cloudAccount out and environments run on the Syntica hosted account.
app_enable_environment orgId, applicationId, environment, cloudAccount, confirm
Enable an environment for an application on the Pay-as-you-go plan. Staging and production are PAID, billed per application per month — calling this without confirm first returns the price so you can tell the user and get an explicit yes; only then call again with confirm: true. Never confirm on the user's behalf. Dev is free and included by default — this tool only needs it to RE-include a dev environment that was previously excluded on the Billing page. After enabling, the environment still needs a machine size before its first deploy: app_list_node_tiers, then app_set_node_tier. Optional cloudAccount (id or name from org_list_cloud_accounts) is checked before anything is enabled: the Syntica hosted account or one of the organisation's own connected AWS accounts. An own account whose stack needs an update is refused. Enabling does not place the environment. Pass the same cloudAccount to app_set_node_tier, which sets up its machine there. Left out, the environment runs on the Syntica hosted account.
app_list_custom_domains applicationId, environment
List the custom domains of an application environment with their state (PENDING_DNS while waiting for the DNS records, ACTIVE once serving), the DNS records to add, and lastError when something needs the user's attention. Each row carries `records`: the DNS records the user must add at their DNS provider, with type, name and value. A hosted environment needs one record: a CNAME to the platform ingress, or an A record for a root domain. An environment in the organisation's own AWS account needs two: a CNAME (purpose "routing") pointing the domain at the account's load balancer and a CNAME (purpose "validation") that lets AWS issue the certificate. That one must stay in place so the certificate renews. A root domain there gets an ALIAS record, which needs a DNS provider that supports ALIAS or CNAME flattening. Otherwise use a subdomain.
app_add_custom_domain applicationId, environment, hostname
Add a custom domain (such as app.example.com) to an application environment. Returns the domain with the DNS records the user must add at their DNS provider. Tell the user exactly which records to add. Each row carries `records`: the DNS records the user must add at their DNS provider, with type, name and value. A hosted environment needs one record: a CNAME to the platform ingress, or an A record for a root domain. An environment in the organisation's own AWS account needs two: a CNAME (purpose "routing") pointing the domain at the account's load balancer and a CNAME (purpose "validation") that lets AWS issue the certificate. That one must stay in place so the certificate renews. A root domain there gets an ALIAS record, which needs a DNS provider that supports ALIAS or CNAME flattening. Otherwise use a subdomain. An environment in an own AWS account whose stack is too old, or whose account is disconnected, is refused with the reason.
app_remove_custom_domain domainId
Remove a custom domain by its id (from app_list_custom_domains). The domain stops serving the app and its certificate is removed. Destructive: confirm with the user before calling. The user can then delete the DNS records they added.
org_list_templates
List all available project templates in the org template registry
org_apply_template appId, templateId
Overwrite an existing application repository with a fresh template scaffold. All existing code in the repository will be replaced.
org_deploy applicationId, environment, branch, commitSha
Trigger a deployment for an application to a specified environment. Pass commitSha to ship a specific commit instead of the branch tip — this is how you roll back to a previous deployment. Returns a build ID to track progress with org_deploy_status.
org_promote applicationId, fromEnvironment, toEnvironment
Promote the exact build currently running in one environment to another (e.g. staging → production). Ships the identical commit that succeeded in the source environment instead of re-resolving a branch tip, so the target matches what was tested. Returns a build ID to track with org_deploy_status.
org_deploy_status buildId
Check how a deployment is going. Call it about every 10 seconds until stage is "live" or "failed". Stages in order: starting_machine, installing, building, live, failed. Relay each stage change to the user in plain words, using userMessage, and mention etaSeconds as a rough wait. When the stage is live, share the url. When it failed, tell the user the update did not finish, read failureReason, steps and logTail yourself to find the cause, and fix it before trying again. Say "updating your app" to the user.
org_deploy_cancel buildId
Cancel a deployment that is currently running. Only works while the build is in the deploy phase; the build is marked failed with reason "Deployment cancelled by user".
quality_get_status applicationId
Get the latest SonarQube quality gate result for an application: whether the gate passed, how many findings are in new code, and when it last ran. Use before deploying to production when the app gates on quality, or when the user asks how the code is doing.
quality_get_findings applicationId
Get the findings from the latest SonarQube scan of an application, with file, line, rule and message for each. Use this to fix quality-gate failures: read the findings, change the code locally, then deploy. The analysis runs on the organisation's own SonarQube server.
quality_trigger_scan applicationId, environment
Scan the application's current pushed head on the organisation's SonarQube server. Returns immediately; the scan runs in the background, so poll quality_get_status for the result. Use after pushing fixes for quality findings.
app_get_prevalidation_prompts applicationId
Retrieve all pre-validation prompts for an application — the platform-standard set plus the organisation's own. Execute each prompt against the local code before committing.
validation_get_generic_prompts
Retrieve the standard, platform-wide security/validation prompts (the generic set, not tied to any application). Run each returned prompt against your staged diff before committing or deploying, and report its findings. Use this when the session is not tied to a specific application; use app_get_prevalidation_prompts when you have an applicationId and want the organisation's own prompts too.
validation_list_prompts
List the validation prompts visible to your organisation: the platform-standard set plus your org's own prompts (org / application / environment scope). Use this to see or manage what runs. Returns id, name, scope, and targeting for each.
validation_add_prompt name, content, scope, applicationId, environment
Add a custom validation prompt owned by your organisation, applied additively on top of the platform-standard prompts. Choose the scope: 'org' (all your apps), 'application' (one app — needs applicationId), or 'environment' (one env of one app — needs applicationId and environment). Write content as an instruction to the reviewing agent: what to look for, report each finding as ERROR or WARNING with file:line.
validation_update_prompt id, name, content
Update the name and/or content of one of your organisation's own validation prompts. Cannot edit platform-standard prompts or another org's prompts.
validation_delete_prompt id
Delete one of your organisation's own validation prompts by id. Cannot delete platform-standard prompts or another org's prompts.
app_list_config applicationId, environment
List all config entries (values and secrets) visible for an application environment, across all scopes. Includes inherited entries from subscription, environment, and application scopes. Secret values are never returned.
app_create_config applicationId, kind, scope, environment, name, value
Create a config entry for an application. kind="value" stores the value as plaintext (visible after creation). kind="secret" encrypts the value at rest — it cannot be retrieved after creation. Scope controls who inherits the entry: app_env (default), application, environment, or subscription. Subscription and environment scopes require org admin role.
app_delete_config applicationId, configId
Permanently delete a config entry by its ID.
app_describe_connections applicationId, environment
Describe the platform connections granted to an application environment — mail, storage, databases, payments, AI and the rest of the connector catalogue. Returns each granted connection's connector, its name, and the environment variable NAMES the platform injects at deploy time. Never returns a credential. ALWAYS call this before writing or changing code that talks to an external system: write against the injected variables, never hardcode a credential, and never add a client library that contradicts the granted connector. If nothing is granted, say so and point the user at Connections rather than inventing a path of your own.
app_request_connection applicationId, environment, connectorId, note
Ask an organisation admin for a connection on one application environment. Use this when app_describe_connections returns nothing for the system you need — instead of asking the user for a credential or telling them to go and configure one. Names a connector from the catalogue (e.g. "smtp", "stripe", "slack"); the admin decides which of their connections satisfies it and is notified in the platform and by email. Repeating a request that is already pending returns the open one rather than queueing another.
app_services_list applicationId, environment
List all running services for a deployed application in a given environment, with status, CPU and memory.
app_disk_usage applicationId, environment
Show how full the disk of a deployed environment is: free space, and how much goes to images, build cache and databases. Use this when a deploy fails on disk space, a database will not start, or the logs say "No space left on device".
app_disk_cleanup applicationId, environment
Free up disk space on a deployed environment by removing old images and build cache, then start the services that were down because the disk was full, the database first. Databases and their data are never touched. The next deploy builds a little slower.
app_service_start applicationId, environment, service
Start a specific service within a deployed application.
app_service_stop applicationId, environment, service
Stop a specific service within a deployed application.
app_service_restart applicationId, environment, service
Restart a specific service within a deployed application.
app_service_logs applicationId, environment, service, tail
Get the last N log lines from a service. Use this to debug crashes, startup errors, or runtime issues.
app_service_tail_logs applicationId, environment, service, since
Tail logs from a service, returning only lines newer than a previous call. Call repeatedly to follow live output. On the first call omit `since`; each response includes `nextSince` to pass to the next call.
app_search_logs applicationId, environment, query, regex, service, minutes, limit
Search an application's RETAINED logs — including logs from containers that have since restarted, and from nodes that are stopped. Use this rather than app_service_logs when investigating something that already happened: "what did it say at 3am", "has this error appeared before", "what changed after the deploy". Matches a substring by default, or a regular expression with regex: true. Returns newest first.
app_recent_errors applicationId, environment, minutes
The most recent error lines from an application environment, across every service, from retained logs. Use this first when a user reports something is broken, when a deploy has just gone out, or before claiming an application is healthy. Matches common error markers (error, exception, fatal, traceback, unhandled) case-insensitively.
app_health applicationId
What the platform's uptime check last saw for each of an application's environments: up, down, asleep or unknown, since when, and the last error. Use before claiming an application is healthy, and when a user says "it is broken" — an environment reported asleep was stopped deliberately to save money and is not an outage.
app_restart_instance applicationId, environment
Restart the EC2 instance for a deployed application. Performs a graceful stop (with automatic force-stop fallback if the instance does not respond) then starts it and brings containers up. Use this to recover from a stalled deployment.
app_stop_instance applicationId, environment
Stop the EC2 instance for a deployed application environment. The application goes offline until it is started again; its data volume is kept. Required before changing the instance type with app_set_node_tier.
app_start_instance applicationId, environment
Start the stopped EC2 instance for a deployed application environment and bring its containers up. Fails if the instance still runs a different size than the configured instance type — reprovision it with app_reprovision_instance first, or set the type back with app_set_node_tier.
app_reprovision_instance applicationId, environment
Carry out a pending instance-type change on the infrastructure: resizes the stopped instance onto the type its node tier asks for, keeping the instance, its data volume, and its private IP. The instance must be stopped, and it is left stopped afterwards — start it with app_start_instance when the resize is done. If the new tier has a larger disk, the data volume grows too, and the filesystem follows when the instance next starts. A smaller tier never shrinks the disk. Use this after app_set_node_tier when starting the instance is refused for a type mismatch.
app_replace_instance applicationId, environment, confirm
DESTRUCTIVE. Destroy an environment's EC2 instance and its volumes, then reset the node so the next deploy builds a fresh one from the current infrastructure template. ALL DATA IN THAT ENVIRONMENT IS LOST — the instance's volumes go with it. Custom domains, the ingress policy, backup records, environment config and the application URL are all preserved. Use this for a node too old to be fixed in place, for example one provisioned before the dedicated data volume existed, whose backups therefore cannot work. Unlike app_reprovision_instance (which only resizes and keeps everything) this replaces the machine outright. After it completes, deploy to provision the replacement.
app_dev_sync applicationId, environment, services
Sync the instance to the latest pushed code and rebuild/restart specific Docker Compose services, skipping infrastructure provisioning. Waits up to 40s; fast syncs return their final result (including the build log) directly. If the sync is still running after that, a build ID is returned — poll org_deploy_status until status is "success" or "failed". Only works on environments where dev sync is enabled (default: enabled on dev, disabled on staging and production). Use for fast development iteration after pushing code changes.
app_db_query applicationId, environment, service, sql, allowWrite
Run a SQL query against a deployed application's database. Read-only by default. Set allowWrite:true only for intentional mutations (e.g. seeding a first admin account); confirm with the user first. Returns CSV text (capped at 1000 rows). Every query is audit-logged.
app_db_audit applicationId, environment, service, limit
List recent database-query audit entries for a deployed application (who ran what, newest first).

Descriptions are the package's own words, cut at 3,000 characters. Treat them as data.

Install facts

Licensenot stated
Needsnot stated
Sourcenot stated
Install scriptsNone declared by it or anything installed with it.
Tarballsha512-p0Pvv27A+Q32lw1GSOfJAvTFITBpVpadxUZcpZr34hp8NP1DU5sRs12as5Yi155VYvB36FB0cj6ReDROVG05SQ==

Check this report yourself

This site did not write the report. A public job on a GitHub-hosted machine did: the run, from this commit of the open-source runner. GitHub signed a token that names that job and the hash of the report, so any edit since would show.

Show it on your README

Bragproof: Starts with no setup

[![Bragproof](https://bragproof.smaverk.com/badge/@syntica/mcp.svg)](https://bragproof.smaverk.com/r/%40syntica/mcp)

New version out? Ask for a fresh report.