Reference

MCP tools

Every tool the Stableship MCP server offers your coding agent, the scope it needs, and the CLI command that does the same.

The server

https://api.stableship.ai/mcp

It speaks MCP over streamable HTTP and needs no local bridge. Clients sign in through the browser, or send a token as Authorization: Bearer sst_your_token. See Sign-in and tokens.

The server tells a connected agent how to use the tools when it connects. Most agents need no more than "fix the blocking Stableship findings and verify each fix".

Scopes

A connection that signs in through the browser can use every tool. A token can use the tools its scopes allow: read tools only read, run tools start work or add to your notes and rules, and manage tools act for you in the app. A tool that would lower protection asks a person first: see Run Stableship from your agent.

Tools

Findings and fixes

Tool	Scope	What it does
`list_projects`	read	Your apps, each with its repositories, latest verdict and counts of open and blocking findings. CLI: `stableship status`.
`get_findings`	read	An app's findings, blocking ones first. Filter by status, blocking only, or first seen since a time. CLI: `stableship findings`.
`get_fix_brief`	read	Everything needed to fix one finding: what is wrong, the rule it breaks, where to look, how to reproduce it, when it is done, and a ready prompt. CLI: `stableship brief <findingId>`.
`explain_finding`	read	Why a finding blocks or does not, or why a run came out as it did. CLI: `stableship why <findingId\|runId>`.
`get_known_issues`	read	Accepted risks, false positives, won't-fix and in-progress items, and dismissed findings. CLI: `stableship known-issues`.
`get_fix_plan`	read	Every open finding as ordered tasks, blocking first: STABLESHIP_FIXES.md. CLI: `stableship fix-plan`.
`verify_fix`	run	Run again the check that found a finding, on a branch or commit, to confirm a fix. CLI: `stableship verify <findingId>`.

Checks and audits

Tool	Scope	What it does
`run_check`	run	Start a check of a pushed commit or branch, at a scope you choose or the one the change needs. Can be narrowed to one flow, kind of user or endpoint. CLI: `stableship check`.
`get_run_status`	read	A run's status and verdict, whether it is done, and its findings.
`rerun_gate`	run	Run a check again with the same scope, narrowing and commit. CLI: `stableship rerun <runId>`.
`run_audit`	run	Start a full audit of the default branch, from the app's monthly audits or a purchased one. CLI: `stableship audit`.
`get_audit_status`	read	Whether a full audit is running or can start, without starting one.
`retest_untested`	run	Test again what the last audit could not test, once the cause is fixed.
`get_scope_policy`	read	The scope minimums and blocking policy, and the scope a check would pick for a set of paths.

Rules, notes and flows

Tool	Scope	What it does
`get_rules`	read	The app's rules and critical flows, or only the rules that concern one file. CLI: `stableship rules`.
`get_context`	read	The rules as sentences, the kinds of user, your product notes, and what the last private copy could not test. CLI: `stableship context`.
`add_context_note`	run	Add one lasting fact to the product notes. CLI: `stableship note "<fact>"`.
`remove_context_note`	run	Take one product note out. CLI: `stableship note --remove "<words>"`.
`propose_rule`	run	Suggest who may do what, for a person to accept, instead of changing access in code. CLI: `stableship propose`.
`add_flow`	run	Add something people do that every check clicks through. CLI: `stableship flow`.
`get_repo_files`	read	The .gate files and the Stableship section of AGENTS.md, for your agent to commit.

Reports and setup

Tool	Scope	What it does
`get_report`	read	The audit report: verdict, steps, problems and what was not tested. CLI: `stableship report`.
`get_launch_report`	read	The latest launch report.
`get_setup`	read	The short block for CLAUDE.md or AGENTS.md that points your agent at the rules and the fix plan. CLI: `stableship setup`.

Your account and app

Tool	Scope	What it does
`whoami`	read	The workspace, the apps this connection can see, its scopes, and whose it is. CLI: `stableship whoami`.
`get_next_steps`	read	The latest audit as a plan: how it went, what to fix first, and the one next step. CLI: `stableship next`.
`get_activity`	read	What changed recently and who did it. CLI: `stableship activity`.
`get_app_map`	read	The pages each kind of user reached and how each went, the endpoints the checks called, and the outside services and personal data the app holds. CLI: `stableship map`.
`get_compliance`	read	How the app maps to SOC 2, ISO 27001, GDPR, PCI DSS and more: what was proven, the gaps, and what to ask a professional.
`get_approval`	read	Whether a person approved an action, and whether Stableship carried it out. CLI: `stableship approval <approvalId>`.

Actions (manage scope)

Tool	Scope	What it does
`triage_finding`	manage	Mark a finding a false positive, an accepted risk, fixed, or open again. CLI: `stableship do triage_finding`.
`resume_held_audit`	manage	Start again an audit that was on hold because the private copy could not be built. CLI: `stableship do resume_held_audit`.
`ship_anyway`	manage	Let a held release go out, with a reason that is kept in the activity log. CLI: `stableship do ship_anyway`.
`decide_rule_suggestion`	manage	Accept or dismiss a rule suggestion. CLI: `stableship do decide_rule_suggestion`.
`set_policy`	manage	Change a policy: what blocks, what happens when Stableship cannot finish, the audit schedule, notifications. CLI: `stableship do set_policy`.
`set_autonomy`	manage	Change how much Stableship does on its own: hands on, guided or autopilot. CLI: `stableship do set_autonomy`.
`set_role`	manage	Add or rename a kind of user. CLI: `stableship do set_role`.
`set_role_reach`	manage	Set whether a kind of user reaches every record or only related ones. CLI: `stableship do set_role_reach`.
`remove_role`	manage	Remove a kind of user, so checks stop testing as it. CLI: `stableship do remove_role`.
`remove_rule`	manage	Remove a rule, so checks stop testing it. Removing an access or money rule needs the approvers the policy names. CLI: `stableship do remove_rule`.
`share_report`	manage	Make an audit report readable by anyone with its link. CLI: `stableship do share_report`.
`unshare_report`	manage	Stop sharing an audit report. CLI: `stableship do unshare_report`.
`reanalyze`	manage	Read the app again: its pages, data and flows. CLI: `stableship do reanalyze`.
`set_notifications`	manage	Choose which emails you get. CLI: `stableship do set_notifications`.
`create_app`	manage	Add an app from one or more repositories. CLI: `stableship do create_app`.
`link_repo`	manage	Add a repository to an app. CLI: `stableship do link_repo`.
`set_service`	manage	Correct how one of the app's services is built and started. CLI: `stableship do set_service`.

For your coding agent: this page as Markdown