Skip to content
Sign up

flagon-io/g1t

public
0
897 linesCodeBlameRaw

Pick any line to see why it is the way it is: the commit, the pull request and issue it came from, and what the agent was thinking.

Thirteen MCP tools and classic token scopes; agents rate their confidence and can be put on an issue in one step1//! The MCP server's tools: a few resource tools, each with an `action`.
2//!
3//! Every operation is one action of one tool. A call is dispatched to the
4//! operation it names, so permissions, the audit log, billing and outcomes
5//! are exactly those of the REST API. A token sees only the actions its
6//! scopes allow, and a tool none of whose actions it may use is not listed.
7//!
8//! The listed input schema is one flat object: `action`, then every field
9//! any of its actions takes. Which fields each action needs is in the
10//! `action` field's description and checked on every call. Claude's API,
11//! and so most MCP clients, refuse a tool whose input schema has `oneOf`
12//! at its top level, so the schema keyed by action, with each action's
13//! required fields, is [`discriminated`], published on the server's card
14//! and in the docs.
15
16use g1t_contracts::credentials::NEVER;
17use g1t_contracts::identity::AgentScope;
18use g1t_contracts::scopes::{Level, NO_SCOPE, TokenAccess, scope_for};
19use serde_json::{Map, Value, json};
20
Deployments wherever they run: API, g1t Actions environments, and a Deployments section21use crate::deployments::DeploymentsOp;
Thirteen MCP tools and classic token scopes; agents rate their confidence and can be put on an issue in one step22use crate::operations::Op;
Merge rulesets: branch and tag rules, agent-first, enforced on push and merge23use crate::rules::RulesOp;
Teams and CODEOWNERS, labels and milestones, dependency updates, the security suite, and a clearer top bar24use crate::security::SecurityOp;
Thirteen MCP tools and classic token scopes; agents rate their confidence and can be put on an issue in one step25
26pub struct Action {
27 pub name: &'static str,
28 pub op: Op,
29 /// One line, for the `action` field's description.
30 pub summary: &'static str,
31}
32
33pub struct Tool {
34 pub name: &'static str,
35 pub title: &'static str,
36 /// What it is for, in a sentence or two.
37 pub description: &'static str,
38 pub actions: &'static [Action],
39 /// The action a call without one runs.
40 pub default_action: Option<&'static str>,
41}
42
43const fn a(name: &'static str, op: Op, summary: &'static str) -> Action {
44 Action { name, op, summary }
45}
46
47pub const TOOLS: &[Tool] = &[
48 Tool {
49 name: "search",
50 title: "Search",
51 description: "Find things. `code` (the default) searches all of g1t you can see: repositories, code, issues, pull requests and people, with qualifiers like repo:owner/name, language:rust, is:issue. `context` searches one workspace's catalog, docs, issues and memory by meaning.",
52 default_action: Some("code"),
53 actions: &[
54 a("code", Op::Search, "Search all of g1t: repositories, code, issues, pull requests, people"),
55 a("context", Op::SearchContext, "Search a workspace's context hub by meaning"),
56 a("entity", Op::GetEntity, "One catalog entry and its relations"),
57 a("ticket", Op::GetContext, "A Jira, Linear or Sentry item the work refers to, as it is now"),
58 ],
59 },
60 Tool {
61 name: "repository",
62 title: "Repositories",
Merge rulesets: branch and tag rules, agent-first, enforced on push and merge63 description: "Repositories: find, read and create them, change their settings and rulesets (what may happen to branches and tags, and what a pull request needs to merge), check their CODEOWNERS file, manage their labels and milestones, and see and dismiss their security alerts (secrets and vulnerable dependencies). Name one as \"owner/name\". Deleting, transferring and changing visibility need `confirm`.",
Thirteen MCP tools and classic token scopes; agents rate their confidence and can be put on an issue in one step64 default_action: None,
65 actions: &[
66 a("list", Op::ListRepos, "Repositories you can see"),
67 a("get", Op::GetRepo, "One repository"),
68 a("create", Op::CreateRepo, "Create one, empty or copied from a public git URL"),
69 a("update", Op::UpdateRepo, "Change description, website, topics, default branch, protection"),
Merge rulesets: branch and tag rules, agent-first, enforced on push and merge70 a("get_settings", Op::GetRepoSettings, "How pull requests merge, and the default branch's protection as its rules stack"),
71 a("update_settings", Op::UpdateRepoSettings, "Change how pull requests merge and the default branch protection ruleset"),
72 a("check_names", Op::ListCheckNames, "Check names reported lately, to require in a ruleset"),
73 a("list_rulesets", Op::Rules(RulesOp::ListRepoRulesets), "Its rulesets, and its workspace's that hold in it"),
74 a("get_ruleset", Op::Rules(RulesOp::GetRepoRuleset), "One ruleset"),
75 a("create_ruleset", Op::Rules(RulesOp::CreateRepoRuleset), "Create a ruleset for its branches or tags"),
76 a("update_ruleset", Op::Rules(RulesOp::UpdateRepoRuleset), "Change a ruleset"),
77 a("delete_ruleset", Op::Rules(RulesOp::DeleteRepoRuleset), "Delete a ruleset"),
78 a("branch_rules", Op::Rules(RulesOp::GetBranchRules), "Every rule that holds for a branch or tag, and where it comes from"),
79 a("rule_evaluations", Op::Rules(RulesOp::ListRuleEvaluations), "How its rules judged pushes and merges, with insights"),
Teams and CODEOWNERS, labels and milestones, dependency updates, the security suite, and a clearer top bar80 a("codeowners", Op::GetCodeownersErrors, "Problems in its CODEOWNERS file, by line"),
81 a("list_labels", Op::ListLabels, "Labels, with colors and how many issues and pull requests carry each"),
82 a("create_label", Op::CreateLabel, "Create a label"),
83 a("update_label", Op::UpdateLabel, "Rename a label or change its color or description"),
84 a("delete_label", Op::DeleteLabel, "Delete a label, from everything that carries it"),
85 a("add_default_labels", Op::AddDefaultLabels, "Add the default labels it is missing"),
86 a("list_milestones", Op::ListMilestones, "Milestones, with progress and due dates"),
87 a("get_milestone", Op::GetMilestone, "One milestone with its issues and pull requests"),
88 a("create_milestone", Op::CreateMilestone, "Create a milestone"),
89 a("update_milestone", Op::UpdateMilestone, "Change a milestone's title, description, due date or state"),
90 a("delete_milestone", Op::DeleteMilestone, "Delete a milestone"),
Thirteen MCP tools and classic token scopes; agents rate their confidence and can be put on an issue in one step91 a("list_events", Op::ListEvents, "Timeline: pushes, issues, pull requests, comments"),
92 a("rename_branch", Op::RenameBranch, "Rename a branch"),
93 a("rename", Op::RenameRepo, "Rename it; old addresses redirect"),
94 a("transfer", Op::TransferRepo, "Move it to another workspace you own"),
95 a("archive", Op::ArchiveRepo, "Make it read-only"),
96 a("unarchive", Op::UnarchiveRepo, "Make it writable again"),
97 a("set_visibility", Op::SetRepoVisibility, "Make it public or private"),
98 a("delete", Op::DeleteRepo, "Delete it; restorable for 30 days"),
99 a("list_deleted", Op::ListDeletedRepos, "A workspace's deleted repositories"),
100 a("restore", Op::RestoreRepo, "Restore a deleted one"),
101 a("purge", Op::PurgeRepo, "Remove a deleted one for good"),
Git storage hardened, pages in tens of milliseconds, honest security alerts, and costs reconciled daily102 a("security_alerts", Op::ListSecurityAlerts, "Secret and dependency alerts, filtered by state"),
103 a("dismiss_alert", Op::DismissSecurityAlert, "Dismiss an alert with a reason"),
104 a("reopen_alert", Op::ReopenSecurityAlert, "Reopen a dismissed alert"),
Thirteen MCP tools and classic token scopes; agents rate their confidence and can be put on an issue in one step105 ],
106 },
107 Tool {
108 name: "issue",
109 title: "Issues",
110 description: "Issues: what should change. Read one before working on it to see the pull requests already made for it. Issues and pull requests share numbers; `comment` works on either.",
111 default_action: None,
112 actions: &[
113 a("list", Op::ListIssues, "Issues on a repository, newest first"),
Fast pages, required checks on the branch, self-hosted runners, honest incidents114 a("get", Op::GetIssue, "One issue with comments and its pull requests"),
Thirteen MCP tools and classic token scopes; agents rate their confidence and can be put on an issue in one step115 a("create", Op::CreateIssue, "Open an issue"),
Teams and CODEOWNERS, labels and milestones, dependency updates, the security suite, and a clearer top bar116 a("update", Op::UpdateIssue, "Change title, body, labels, milestone or assignees"),
117 a("labels", Op::ListIssueLabels, "The labels an issue or pull request carries"),
118 a("add_labels", Op::AddIssueLabels, "Add labels to an issue or pull request"),
119 a("set_labels", Op::SetIssueLabels, "Replace the labels of an issue or pull request"),
120 a("remove_labels", Op::RemoveIssueLabels, "Take labels off an issue or pull request"),
Thirteen MCP tools and classic token scopes; agents rate their confidence and can be put on an issue in one step121 a("close", Op::CloseIssue, "Close it without a pull request"),
122 a("reopen", Op::ReopenIssue, "Reopen it"),
123 a("comment", Op::AddComment, "Comment on an issue or pull request; path and line for one line of a change"),
124 a("import", Op::ImportIssue, "Open an issue from a Jira, Linear or Sentry item"),
125 ],
126 },
127 Tool {
128 name: "pull_request",
129 title: "Pull requests",
Teams and CODEOWNERS, labels and milestones, dependency updates, the security suite, and a clearer top bar130 description: "Pull requests: start a change for an issue, record your session, mark it ready, ask people and teams to review, review and merge. Read `overlaps` and `behind` on `get` before going far, and `code_owners` for whose approval it needs.",
Thirteen MCP tools and classic token scopes; agents rate their confidence and can be put on an issue in one step131 default_action: None,
132 actions: &[
133 a("list", Op::ListPullRequests, "Pull requests on a repository, newest first"),
Fast pages, required checks on the branch, self-hosted runners, honest incidents134 a("get", Op::GetPullRequest, "Status, checks and required checks, reviews, overlaps, whether it is behind"),
Thirteen MCP tools and classic token scopes; agents rate their confidence and can be put on an issue in one step135 a("changes", Op::GetPullRequestChanges, "Files and line-by-line diff"),
136 a("create", Op::CreatePullRequest, "Start a draft with its own fork to push to, or open one from a pushed branch"),
Teams and CODEOWNERS, labels and milestones, dependency updates, the security suite, and a clearer top bar137 a("update", Op::UpdatePullRequest, "Change its base branch, labels, milestone, assignees or reviewers"),
Thirteen MCP tools and classic token scopes; agents rate their confidence and can be put on an issue in one step138 a("record_session", Op::RecordSession, "Append prompt, reasoning and tool entries to its session"),
139 a("read_session", Op::ReadSession, "Its recorded session"),
140 a("ready", Op::MarkPullRequestReady, "Mark a draft ready, with a summary"),
Teams and CODEOWNERS, labels and milestones, dependency updates, the security suite, and a clearer top bar141 a("request_reviewers", Op::RequestReviewers, "Ask people or teams to review it"),
142 a("remove_requested_reviewers", Op::RemoveRequestedReviewers, "Stop asking people or teams to review it"),
Thirteen MCP tools and classic token scopes; agents rate their confidence and can be put on an issue in one step143 a("review", Op::ReviewPullRequest, "Approve or request changes"),
144 a("close", Op::ClosePullRequest, "Close without merging"),
145 a("merge", Op::MergePullRequest, "Land it, or join the merge queue"),
146 a("merge_queue", Op::GetMergeQueue, "The repository's merge queue"),
147 ],
148 },
149 Tool {
150 name: "agent",
151 title: "g1t agents",
152 description: "Put g1t's agent to work and talk to it. One agent per issue; to do more at once, use more issues. Starting an agent uses the workspace's money.",
153 default_action: None,
154 actions: &[
155 a("delegate", Op::Delegate, "Open an issue and put an agent on it in one step"),
156 a("assign", Op::AssignIssue, "Put an agent on an existing issue"),
157 a("message", Op::MessageAgent, "Tell the agent on a pull request something, or ask another agent"),
158 a("answer", Op::AnswerMessage, "Answer a question or handoff sent to you"),
159 a("take_messages", Op::TakeMessages, "For a g1t agent: messages not seen yet"),
160 ],
161 },
162 Tool {
163 name: "plan",
164 title: "Plans",
Fast pages, required checks on the branch, self-hosted runners, honest incidents165 description: "Turn an outcome into issues: an agent proposes them with what done means and their dependencies; nothing opens until you apply the plan.",
Thirteen MCP tools and classic token scopes; agents rate their confidence and can be put on an issue in one step166 default_action: None,
167 actions: &[
168 a("create", Op::PlanWork, "Ask an agent for a plan; read it with get until ready"),
169 a("get", Op::GetPlan, "A plan and the issues it proposes"),
170 a("apply", Op::ApplyPlan, "Open its issues; with assign, agents start in dependency order"),
171 ],
172 },
173 Tool {
174 name: "memory",
175 title: "Memory",
176 description: "What the project and its workspace remember for the next agent: how to build, conventions, decisions, traps. Recall before you start; remember one short fact at a time, never a secret.",
177 default_action: None,
178 actions: &[
179 a("recall", Op::Recall, "Search memory, or list it all"),
180 a("remember", Op::Remember, "Save one fact"),
181 ],
182 },
183 Tool {
184 name: "workflow",
185 title: "Workflows",
Deployments wherever they run: API, g1t Actions environments, and a Deployments section186 description: "GitHub Actions workflows from .g1t/workflows: their runs, jobs and logs, and running, cancelling or rerunning them. Deployments wherever they run (reported from any CI, made by jobs with an `environment:`, or built on g1t.page), their statuses and environments, and reporting your own. Also the self-hosted runners they run on: a workspace's (`workspace`) or a repository's own (`repo`), their groups, and where agent work runs.",
Thirteen MCP tools and classic token scopes; agents rate their confidence and can be put on an issue in one step187 default_action: None,
188 actions: &[
189 a("list", Op::ListWorkflows, "Workflows on the default branch"),
190 a("list_runs", Op::ListWorkflowRuns, "Runs, newest first"),
191 a("get_run", Op::GetWorkflowRun, "One run with its jobs and steps"),
192 a("job_logs", Op::GetJobLogs, "A job's log after a sequence number"),
193 a("dispatch", Op::DispatchWorkflow, "Run a workflow_dispatch workflow"),
194 a("cancel", Op::CancelWorkflowRun, "Cancel a run"),
195 a("rerun", Op::RerunWorkflowRun, "Run a finished run again"),
196 a("update", Op::UpdateWorkflow, "Turn a workflow on or off"),
Deployments wherever they run: API, g1t Actions environments, and a Deployments section197 a("list_deployments", Op::Deployments(DeploymentsOp::ListDeployments), "Deployments wherever they run, newest first, filtered"),
198 a("get_deployment", Op::Deployments(DeploymentsOp::GetDeployment), "One deployment with every status it has had"),
199 a("create_deployment", Op::Deployments(DeploymentsOp::CreateDeployment), "Report a deployment of a ref to an environment"),
200 a("deployment_statuses", Op::Deployments(DeploymentsOp::ListDeploymentStatuses), "A deployment's statuses, newest first"),
201 a("create_deployment_status", Op::Deployments(DeploymentsOp::CreateDeploymentStatus), "Report where a deployment is: in_progress, success, failure"),
202 a("list_environments", Op::Deployments(DeploymentsOp::ListEnvironments), "Environments with their current and latest deployments"),
203 a("get_environment", Op::Deployments(DeploymentsOp::GetEnvironment), "One environment by name"),
Fast pages, required checks on the branch, self-hosted runners, honest incidents204 a("list_runners", Op::ListRunners, "Self-hosted runners, with status, labels and what each is doing"),
205 a("create_runner_token", Op::CreateRunnerRegistrationToken, "A one-hour token for g1t-runner register"),
206 a("remove_runner", Op::RemoveRunner, "Remove a self-hosted runner"),
207 a("list_runner_groups", Op::ListRunnerGroups, "A workspace's runner groups"),
208 a("create_runner_group", Op::CreateRunnerGroup, "Make a group, for some repositories"),
209 a("update_runner_group", Op::UpdateRunnerGroup, "Rename a group or change its repositories"),
210 a("delete_runner_group", Op::DeleteRunnerGroup, "Delete a group; its runners join the default"),
211 a("get_runner_settings", Op::GetRunnerSettings, "Where agent work runs; whether forks may use runners"),
212 a("update_runner_settings", Op::UpdateRunnerSettings, "Change them"),
Thirteen MCP tools and classic token scopes; agents rate their confidence and can be put on an issue in one step213 ],
214 },
215 Tool {
216 name: "secret",
217 title: "Secrets and variables",
218 description: "A repository's or workspace's secrets and variables, read by workflows and deployments. Secret values are never returned.",
219 default_action: None,
220 actions: &[
221 a("list_secrets", Op::ListActionsSecrets, "Secrets, without values"),
222 a("set_secret", Op::SetActionsSecret, "Add or change a secret"),
223 a("delete_secret", Op::DeleteActionsSecret, "Remove a secret"),
224 a("list_variables", Op::ListActionsVariables, "Variables, with values"),
225 a("set_variable", Op::SetActionsVariable, "Add or change a variable"),
226 a("delete_variable", Op::DeleteActionsVariable, "Remove a variable"),
227 ],
228 },
229 Tool {
230 name: "webhook",
231 title: "Webhooks",
232 description: "HTTPS addresses sent signed events as they happen, for a repository or a whole workspace.",
233 default_action: None,
234 actions: &[
235 a("list", Op::ListWebhooks, "Webhooks, without secrets"),
236 a("create", Op::CreateWebhook, "Register one; a ping is sent"),
237 a("update", Op::UpdateWebhook, "Change address, events or active"),
238 a("delete", Op::DeleteWebhook, "Remove one"),
239 a("ping", Op::PingWebhook, "Send a ping"),
240 a("list_deliveries", Op::ListWebhookDeliveries, "Latest deliveries"),
241 a("redeliver", Op::RedeliverWebhook, "Send a delivery again"),
242 ],
243 },
244 Tool {
245 name: "access",
246 title: "Who has access",
247 description: "Who has access to a repository and with which role (read, triage, write, maintain, admin), outside collaborators, and a workspace's base permission.",
248 default_action: None,
249 actions: &[
250 a("list_collaborators", Op::ListCollaborators, "Everyone with a role, and pending invitations"),
251 a("get_permission", Op::GetCollaboratorPermission, "One person's role and capabilities"),
252 a("add_collaborator", Op::AddCollaborator, "Give someone a role, by username or email"),
253 a("update_collaborator", Op::UpdateCollaborator, "Change a direct role"),
254 a("remove_collaborator", Op::RemoveCollaborator, "Take away a direct role"),
255 a("list_invitations", Op::ListRepoInvitations, "Pending invitations to a repository"),
256 a("revoke_invitation", Op::RevokeRepoInvitation, "Withdraw one"),
257 a("set_base_permission", Op::SetBasePermission, "What every member gets on each repository"),
258 a("list_outside_collaborators", Op::ListOutsideCollaborators, "People with roles who are not members"),
259 ],
260 },
261 Tool {
Teams and CODEOWNERS, labels and milestones, dependency updates, the security suite, and a clearer top bar262 name: "team",
263 title: "Teams",
264 description: "Teams: groups of a workspace's members, given roles on repositories together, mentioned as @workspace/team and asked to review together. Name one by `workspace` and its slug (`team`). Any member may create a team; the workspace's owners and the team's maintainers manage it. A secret team is seen only by its people and the owners.",
265 default_action: None,
266 actions: &[
267 a("list", Op::ListTeams, "A workspace's teams you can see"),
268 a("get", Op::GetTeam, "One team"),
269 a("create", Op::CreateTeam, "Create a team; you become its maintainer"),
270 a("update", Op::UpdateTeam, "Change its name, slug, description, visibility, parent or notifications"),
271 a("delete", Op::DeleteTeam, "Delete it; its child teams move up"),
272 a("list_members", Op::ListTeamMembers, "Its people and their roles, child teams' with include_child_teams"),
273 a("set_member", Op::SetTeamMember, "Add a member of the workspace, or change their role"),
274 a("remove_member", Op::RemoveTeamMember, "Take someone out of it"),
275 a("list_child_teams", Op::ListChildTeams, "The teams nested under it"),
276 a("list_repos", Op::ListTeamRepos, "The repositories it has a role on"),
277 a("set_repo", Op::SetTeamRepo, "Give it a role on a repository"),
278 a("remove_repo", Op::RemoveTeamRepo, "Take its role on a repository away"),
279 a("set_review_assignment", Op::SetTeamReviewAssignment, "Whom it picks when asked to review"),
280 a("list_user_teams", Op::ListUserTeams, "The teams someone is in"),
281 ],
282 },
283 Tool {
Thirteen MCP tools and classic token scopes; agents rate their confidence and can be put on an issue in one step284 name: "workspace",
285 title: "Workspaces",
Merge rulesets: branch and tag rules, agent-first, enforced on push and merge286 description: "Workspaces own repositories (g1t.sh/{workspace}/{repo}): create, update or delete one, invite members, connect integrations and model providers, set rulesets that hold across its repositories, and keep your own pinned projects at the top of its sidebar.",
Thirteen MCP tools and classic token scopes; agents rate their confidence and can be put on an issue in one step287 default_action: None,
288 actions: &[
Merge branch 'worktree-agent-ad7c6d88d93adc817'289 a("get", Op::GetWorkspace, "A workspace's details and settings"),
Thirteen MCP tools and classic token scopes; agents rate their confidence and can be put on an issue in one step290 a("create", Op::CreateWorkspace, "Create a workspace"),
Packages, with a container registry on g1t.sh; workspaces deleted whole and kept 30 days; Members for every member291 a("delete", Op::DeleteWorkspace, "Delete a workspace and everything in it (support can restore it for 30 days)"),
Merge branch 'worktree-agent-ad7c6d88d93adc817'292 a("update", Op::UpdateWorkspace, "Change its name, description, base permission or who may create teams"),
Thirteen MCP tools and classic token scopes; agents rate their confidence and can be put on an issue in one step293 a("list_invites", Op::ListWorkspaceInvites, "Its invites"),
294 a("invite_member", Op::InviteMember, "Invite an email address"),
295 a("revoke_invite", Op::RevokeWorkspaceInvite, "Revoke a pending invite"),
296 a("list_integrations", Op::ListIntegrations, "Model providers, alert sources, trackers"),
297 a("connect_integration", Op::ConnectIntegration, "Connect one"),
298 a("disconnect_integration", Op::DisconnectIntegration, "Remove one"),
299 a("test_integration", Op::TestIntegration, "Check its credentials"),
300 a("get_model_routes", Op::GetModelRoutes, "Where each kind of work's model requests go"),
301 a("set_model_routes", Op::SetModelRoutes, "Replace them"),
API: pinned projects over REST and MCP302 a("list_pinned_projects", Op::ListPinnedProjects, "Your pinned projects in it, in your order"),
303 a("pin_project", Op::PinProject, "Pin a project, at a position or the end"),
304 a("unpin_project", Op::UnpinProject, "Unpin a project"),
305 a("reorder_pinned_projects", Op::ReorderPinnedProjects, "Put your pins in a new order"),
Merge rulesets: branch and tag rules, agent-first, enforced on push and merge306 a("list_rulesets", Op::Rules(RulesOp::ListWorkspaceRulesets), "Its rulesets, which hold across its repositories"),
307 a("get_ruleset", Op::Rules(RulesOp::GetWorkspaceRuleset), "One of its rulesets"),
308 a("create_ruleset", Op::Rules(RulesOp::CreateWorkspaceRuleset), "Create a ruleset for some or all of its repositories"),
309 a("update_ruleset", Op::Rules(RulesOp::UpdateWorkspaceRuleset), "Change one of its rulesets"),
310 a("delete_ruleset", Op::Rules(RulesOp::DeleteWorkspaceRuleset), "Delete one of its rulesets"),
311 a("rule_evaluations", Op::Rules(RulesOp::ListWorkspaceRuleEvaluations), "How rules judged changes across its repositories"),
Thirteen MCP tools and classic token scopes; agents rate their confidence and can be put on an issue in one step312 ],
313 },
314 Tool {
Usage, Billing settings and prepaid AI credit; fixes from the UX audit315 name: "billing",
316 title: "Billing",
Merge the AI Gateway: Anthropic's Messages API on a workspace's tokens317 description: "A workspace's billing: its usage by product, project and day, its budget (the monthly spend limit, alerts and whether usage pauses at it), its AI credit, its invoices, and its AI Gateway requests. Amounts are whole millionths of a dollar (`_micros`), or cents (`_cents`) where named. Members read it; changing the budget and buying credit are for owners, as people, and never for g1t's agents.",
Usage, Billing settings and prepaid AI credit; fixes from the UX audit318 default_action: Some("usage"),
319 actions: &[
320 a("usage", Op::GetUsage, "Usage over a range of days, by product, meter, project and day, and what paid for it"),
321 a("budget", Op::GetBudget, "The monthly spend limit, what was spent, alerts and whether usage pauses at the limit"),
322 a("set_budget", Op::SetBudget, "Change the spend limit, alerts, pausing or the alert webhook"),
323 a("ai_credit", Op::GetAiCredit, "AI credit left, its grants, auto-reload and how to buy more"),
324 a("buy_ai_credit", Op::BuyAiCredit, "A payment page to buy AI credit, for a person to open"),
325 a("invoices", Op::ListInvoices, "Every invoice, the itemised usage invoices, and the next one so far"),
326 a("billing_details", Op::GetBillingDetails, "Who invoices are made out to and the payment method on file"),
Merge the AI Gateway: Anthropic's Messages API on a workspace's tokens327 a("gateway_requests", Op::ListGatewayRequests, "Recent AI Gateway requests: model, tokens, cost, status and token"),
Usage, Billing settings and prepaid AI credit; fixes from the UX audit328 ],
329 },
330 Tool {
Teams and CODEOWNERS, labels and milestones, dependency updates, the security suite, and a clearer top bar331 name: "security",
332 title: "Security",
333 description: "A repository's security: secret scanning alerts and push protection bypasses, custom secret patterns, code scanning alerts and SARIF uploads, vulnerability alerts, the dependency graph and its SBOM, dependency review, settings, and a workspace's overview. Fix an alert with g1t. Findings are shown to those who can change the code only. Give `repo` (owner/name), or `workspace` for lists across one.",
334 default_action: Some("secret_alerts"),
335 actions: &[
336 a("secret_alerts", Op::Security(SecurityOp::ListSecretAlerts), "Secret scanning alerts; by state, secret_type, validity, bypassed"),
337 a("secret_alert", Op::Security(SecurityOp::GetSecretAlert), "One secret alert, with where it was found and its bypass requests"),
338 a("update_secret_alert", Op::Security(SecurityOp::UpdateSecretAlert), "Dismiss a secret alert with a reason, or reopen it"),
339 a("secret_locations", Op::Security(SecurityOp::ListSecretLocations), "Every file, line and commit a secret is in"),
340 a("bypass", Op::Security(SecurityOp::BypassPushProtection), "Push past push protection with a reason, or ask to"),
341 a("check_validity", Op::Security(SecurityOp::CheckSecretValidity), "Ask a secret's issuer whether it still works"),
342 a("bypass_requests", Op::Security(SecurityOp::ListBypassRequests), "A workspace's push protection bypass requests"),
343 a("review_bypass", Op::Security(SecurityOp::ReviewBypassRequest), "Approve, deny or cancel a bypass request"),
344 a("patterns", Op::Security(SecurityOp::ListCustomPatterns), "Custom secret patterns of a repository or workspace"),
345 a("create_pattern", Op::Security(SecurityOp::CreateCustomPattern), "Create a custom secret pattern, as a draft or published"),
346 a("update_pattern", Op::Security(SecurityOp::UpdateCustomPattern), "Change, publish or unpublish a custom pattern"),
347 a("delete_pattern", Op::Security(SecurityOp::DeleteCustomPattern), "Delete a custom pattern"),
348 a("dry_run_pattern", Op::Security(SecurityOp::DryRunCustomPattern), "Run a pattern over the default branch without saving it"),
349 a("code_alerts", Op::Security(SecurityOp::ListCodeAlerts), "Code scanning alerts; by state, severity, tool, rule_id"),
350 a("code_alert", Op::Security(SecurityOp::GetCodeAlert), "One code scanning alert by number"),
351 a("update_code_alert", Op::Security(SecurityOp::UpdateCodeAlert), "Dismiss a code scanning alert with a reason, or reopen it"),
352 a("analyses", Op::Security(SecurityOp::ListAnalyses), "Code scanning analyses, newest first"),
353 a("upload_sarif", Op::Security(SecurityOp::UploadSarif), "Upload a SARIF file, gzipped and base64-encoded"),
354 a("sarif_upload", Op::Security(SecurityOp::GetSarifUpload), "Whether a SARIF upload was read, and its analyses"),
355 a("vulnerability_alerts", Op::Security(SecurityOp::ListVulnerabilityAlerts), "Vulnerable dependencies; by state, severity, ecosystem, package"),
356 a("vulnerability_alert", Op::Security(SecurityOp::GetVulnerabilityAlert), "One vulnerability alert"),
357 a("update_vulnerability_alert", Op::Security(SecurityOp::UpdateVulnerabilityAlert), "Dismiss a vulnerability alert with a reason, or reopen it"),
358 a("fix", Op::Security(SecurityOp::FixAlert), "Put g1t on an issue to fix an alert"),
359 a("dependency_graph", Op::Security(SecurityOp::GetDependencyGraph), "Every package the lockfiles resolve, direct or transitive"),
360 a("sbom", Op::Security(SecurityOp::GetSbom), "The dependency graph as an SPDX 2.3 document"),
361 a("compare_dependencies", Op::Security(SecurityOp::CompareDependencies), "What changes in dependencies between base...head"),
362 a("settings", Op::Security(SecurityOp::GetSettings), "A repository's security settings"),
363 a("update_settings", Op::Security(SecurityOp::UpdateSettings), "Change when checks fail and dependency review's policy"),
364 a("workspace_settings", Op::Security(SecurityOp::GetWorkspaceSettings), "A workspace's delegated bypass and validity checks"),
365 a("update_workspace_settings", Op::Security(SecurityOp::UpdateWorkspaceSettings), "Turn delegated bypass or validity checks on or off"),
366 a("overview", Op::Security(SecurityOp::GetOverview), "A workspace's alerts, trends and coverage"),
367 ],
368 },
369 Tool {
API: notifications over REST and MCP, with notifications scopes370 name: "notifications",
371 title: "Notifications",
372 description: "Your inbox: what needs you, and what you follow. One thread per issue, pull request, workflow or deployment, with why you were told (`reason`): an agent waiting on you, a review asked of you, an assignment, a mention, your work's checks, or what you subscribe to and watch. Mark threads read or done once handled, and choose what you hear of with subscribe, unsubscribe and watch. Your own: a personal token.",
373 default_action: Some("list"),
374 actions: &[
375 a("list", Op::ListNotifications, "Unread threads, latest first; all, a view, a reason, a repository"),
376 a("get", Op::GetNotificationThread, "One thread with its recent activity and your subscription"),
377 a("mark_read", Op::MarkThreadRead, "Mark a thread read, or unread"),
378 a("mark_all_read", Op::MarkNotificationsRead, "Mark everything read up to a time, or one repository's"),
379 a("done", Op::MarkThreadDone, "Mark a thread done; new activity brings it back"),
380 a("save", Op::SaveThread, "Save a thread, or unsave it"),
381 a("snooze", Op::SnoozeThread, "Snooze a thread until a time, or bring it back"),
382 a("subscription", Op::GetThreadSubscription, "Your subscription to an issue or pull request"),
383 a("subscribe", Op::SetThreadSubscription, "Subscribe to an issue or pull request, or ignore it"),
384 a("unsubscribe", Op::DeleteThreadSubscription, "Unsubscribe until you comment or are mentioned"),
385 a("watching", Op::GetRepoSubscription, "How you watch a repository"),
386 a("watch", Op::SetRepoSubscription, "Watch a repository: participating, all, ignore or custom"),
387 a("unwatch", Op::DeleteRepoSubscription, "Stop watching a repository"),
388 a("watched", Op::ListWatchedRepos, "Repositories you watch other than the default way"),
389 ],
390 },
391 Tool {
Thirteen MCP tools and classic token scopes; agents rate their confidence and can be put on an issue in one step392 name: "account",
393 title: "Your account",
394 description: "Who this token acts as and its workspaces (`whoami`), your email addresses, your invites, and invitations to repositories waiting for you.",
395 default_action: Some("whoami"),
396 actions: &[
397 a("whoami", Op::Whoami, "Who the token acts as, and its workspaces"),
398 a("list_emails", Op::ListEmails, "Your addresses"),
399 a("add_email", Op::AddEmail, "Add an address"),
400 a("remove_email", Op::RemoveEmail, "Remove an address"),
401 a("update_email_settings", Op::UpdateEmailSettings, "Primary, backup and privacy"),
402 a("list_invites", Op::ListInvites, "Your invites to g1t"),
403 a("create_invite", Op::CreateInvite, "Make an invite"),
404 a("revoke_invite", Op::RevokeInvite, "Revoke one"),
405 a("list_repository_invitations", Op::ListMyRepoInvitations, "Invitations to repositories for you"),
406 a("accept_repository_invitation", Op::AcceptRepoInvitation, "Accept one"),
407 a("decline_repository_invitation", Op::DeclineRepoInvitation, "Decline one"),
408 ],
409 },
410];
411
412/// Operations that cannot be undone, or reach beyond g1t's own records:
413/// clients ask before running a tool that has any of them.
414fn destructive(op: Op) -> bool {
415 matches!(
416 op,
Teams and CODEOWNERS, labels and milestones, dependency updates, the security suite, and a clearer top bar417 Op::Security(SecurityOp::DeleteCustomPattern | SecurityOp::BypassPushProtection)
Merge rulesets: branch and tag rules, agent-first, enforced on push and merge418 | Op::Rules(RulesOp::DeleteRepoRuleset | RulesOp::DeleteWorkspaceRuleset)
Teams and CODEOWNERS, labels and milestones, dependency updates, the security suite, and a clearer top bar419 | Op::DeleteWorkspace
Git storage hardened, pages in tens of milliseconds, honest security alerts, and costs reconciled daily420 | Op::UpdateWorkspace
Thirteen MCP tools and classic token scopes; agents rate their confidence and can be put on an issue in one step421 | Op::DeleteRepo
422 | Op::PurgeRepo
423 | Op::TransferRepo
424 | Op::SetRepoVisibility
425 | Op::RemoveEmail
426 | Op::RemoveCollaborator
427 | Op::DisconnectIntegration
428 | Op::DeleteWebhook
429 | Op::DeleteActionsSecret
430 | Op::DeleteActionsVariable
431 | Op::SetActionsSecret
432 | Op::SetActionsVariable
433 | Op::SetModelRoutes
434 | Op::SetBasePermission
Teams and CODEOWNERS, labels and milestones, dependency updates, the security suite, and a clearer top bar435 | Op::DeleteTeam
436 | Op::RemoveTeamRepo
Thirteen MCP tools and classic token scopes; agents rate their confidence and can be put on an issue in one step437 | Op::MergePullRequest
Fast pages, required checks on the branch, self-hosted runners, honest incidents438 | Op::RemoveRunner
439 | Op::DeleteRunnerGroup
440 | Op::UpdateRunnerSettings
Thirteen MCP tools and classic token scopes; agents rate their confidence and can be put on an issue in one step441 )
442}
443
444/// Whether an operation only reads.
445pub fn reads_only(op: Op) -> bool {
446 NO_SCOPE.contains(&op.name())
447 || scope_for(op.name()).is_some_and(|scope| scope.level() == Level::Read)
448}
449
450/// What decides which actions a caller sees.
451pub enum Gate<'a> {
452 /// No limit beyond the person's own role.
453 Everything,
454 /// A g1t agent's token: the operations its run lists.
455 Agent(&'a AgentScope),
456 /// An access token with scopes.
457 Token(&'a TokenAccess),
458}
459
460impl Gate<'_> {
461 pub fn allows(&self, op: Op) -> bool {
462 match self {
463 Gate::Everything => true,
464 Gate::Agent(scope) => op.allowed_by(scope) && !NEVER.contains(&op.name()),
465 Gate::Token(access) => {
466 if NO_SCOPE.contains(&op.name()) {
467 return true;
468 }
469 match scope_for(op.name()) {
470 Some(scope) => access.allows(scope),
471 None => access.scopes.is_none(),
472 }
473 }
474 }
475 }
476}
477
478impl Tool {
479 pub fn by_name(name: &str) -> Option<&'static Tool> {
480 TOOLS.iter().find(|tool| tool.name == name)
481 }
482
483 pub fn action(&self, name: &str) -> Option<&'static Action> {
484 // The tools are 'static; find through TOOLS to keep the lifetime.
485 TOOLS
486 .iter()
487 .find(|tool| tool.name == self.name)
488 .and_then(|tool| tool.actions.iter().find(|action| action.name == name))
489 }
490
491 pub fn visible(&self, gate: &Gate) -> Vec<&'static Action> {
492 TOOLS
493 .iter()
494 .find(|tool| tool.name == self.name)
495 .map(|tool| tool.actions.iter().filter(|action| gate.allows(action.op)).collect())
496 .unwrap_or_default()
497 }
498
499 /// The flat input schema of the actions given.
500 pub fn input_schema(&self, actions: &[&Action]) -> Value {
501 let mut properties = Map::new();
502 let lines: Vec<String> = actions
503 .iter()
504 .map(|action| {
505 let required: Vec<String> = action.op.required();
506 if required.is_empty() {
507 format!("{}: {}.", action.name, action.summary)
508 } else {
509 format!("{} ({}): {}.", action.name, required.join(", "), action.summary)
510 }
511 })
512 .collect();
513 let mut action_schema = json!({
514 "type": "string",
515 "enum": actions.iter().map(|action| action.name).collect::<Vec<_>>(),
516 "description": lines.join("\n"),
517 });
518 if let Some(default) = self.default_action.filter(|name| actions.iter().any(|action| action.name == *name)) {
519 action_schema["default"] = json!(default);
520 }
521 properties.insert("action".to_owned(), action_schema);
522 for action in actions {
523 for (name, schema) in action.op.properties() {
524 merge_property(&mut properties, name, schema);
525 }
526 }
527 let mut required = vec![];
528 if self.default_action.is_none() {
529 required.push("action");
530 }
531 let mut schema = json!({ "type": "object", "properties": properties });
532 if !required.is_empty() {
533 schema["required"] = json!(required);
534 }
535 schema
536 }
537
538 /// The input schema keyed by action: one `oneOf` branch per action,
539 /// each with its own fields and the ones it needs.
540 pub fn discriminated(&self, actions: &[&Action]) -> Value {
541 let branches: Vec<Value> = actions
542 .iter()
543 .map(|action| {
544 let mut properties = Map::new();
545 properties.insert("action".to_owned(), json!({ "const": action.name }));
546 properties.extend(action.op.properties());
547 let mut required = vec![Value::String("action".to_owned())];
548 // The default action may leave `action` out.
549 if self.default_action == Some(action.name) {
550 required.clear();
551 }
552 required.extend(action.op.required().into_iter().map(Value::String));
553 json!({
554 "title": action.name,
555 "description": action.summary,
556 "type": "object",
557 "properties": properties,
558 "required": required,
559 })
560 })
561 .collect();
562 json!({ "type": "object", "oneOf": branches })
563 }
564
565 /// MCP's hints about the actions given: whether the tool only reads,
566 /// whether it can destroy something, and whether calling it twice is
567 /// the same as once.
568 pub fn annotations(&self, actions: &[&Action]) -> Value {
569 let read_only = actions.iter().all(|action| reads_only(action.op));
570 json!({
571 "title": self.title,
572 "readOnlyHint": read_only,
573 "destructiveHint": !read_only && actions.iter().any(|action| destructive(action.op)),
574 "idempotentHint": read_only,
575 "openWorldHint": false,
576 })
577 }
578
579 /// The tool as `tools/list` gives it, for a caller behind `gate`, or
580 /// `None` when it may use none of its actions.
581 pub fn listed(&self, gate: &Gate) -> Option<Value> {
582 let actions = self.visible(gate);
583 if actions.is_empty() {
584 return None;
585 }
586 Some(json!({
587 "name": self.name,
588 "title": self.title,
589 "description": self.description,
590 "inputSchema": self.input_schema(&actions),
591 "annotations": self.annotations(&actions),
592 }))
593 }
594}
595
596/// Adds a property to a tool's flat schema. The first action to use a name
597/// describes it; a later one with other allowed values adds them.
598fn merge_property(properties: &mut Map<String, Value>, name: String, schema: Value) {
599 match properties.get_mut(&name) {
600 None => {
601 properties.insert(name, schema);
602 }
603 Some(existing) => {
604 if let (Some(Value::Array(had)), Some(Value::Array(more))) =
605 (existing.get("enum").cloned(), schema.get("enum"))
606 {
607 let mut merged = had;
608 for value in more {
609 if !merged.contains(value) {
610 merged.push(value.clone());
611 }
612 }
613 existing["enum"] = Value::Array(merged);
614 }
615 // Different kinds of value under one name: say less, accept both.
616 if existing.get("type") != schema.get("type")
617 && let Some(fields) = existing.as_object_mut()
618 {
619 fields.remove("type");
620 fields.remove("items");
621 }
622 }
623 }
624}
625
626/// What a call to a tool runs: the operation its action names, or why not.
627pub fn resolve(tool: &Tool, arguments: &Value) -> Result<Op, String> {
628 let names = || {
629 tool.actions
630 .iter()
631 .map(|action| action.name)
632 .collect::<Vec<_>>()
633 .join(", ")
634 };
635 let Some(name) = arguments["action"].as_str().or(tool.default_action) else {
636 return Err(format!("Give an action: one of {}.", names()));
637 };
638 let Some(action) = tool.action(name) else {
639 return Err(format!("{} has no action {name}. Its actions: {}.", tool.name, names()));
640 };
641 let missing: Vec<String> = action
642 .op
643 .required()
644 .into_iter()
645 .filter(|field| arguments.get(field).is_none_or(Value::is_null))
646 .collect();
647 if !missing.is_empty() {
648 return Err(format!("{}.{name} needs {}.", tool.name, missing.join(", ")));
649 }
650 Ok(action.op)
651}
652
653#[cfg(test)]
654mod tests {
655 use super::*;
656 use g1t_contracts::scopes::{Preset, Scope};
657
658 fn listed(gate: &Gate) -> Vec<Value> {
659 TOOLS.iter().filter_map(|tool| tool.listed(gate)).collect()
660 }
661
662 fn token(scopes: Option<Vec<Scope>>) -> TokenAccess {
663 TokenAccess {
664 token_id: "tok_1".to_owned(),
665 scopes: scopes.map(|scopes| scopes.iter().map(|scope| scope.as_str().to_owned()).collect()),
666 legacy: false,
Merge the AI Gateway: Anthropic's Messages API on a workspace's tokens667 name: None,
Thirteen MCP tools and classic token scopes; agents rate their confidence and can be put on an issue in one step668 }
669 }
670
671 #[test]
672 fn every_operation_is_exactly_one_action_of_one_tool() {
673 for op in Op::ALL {
674 let count = TOOLS
675 .iter()
676 .flat_map(|tool| tool.actions.iter())
677 .filter(|action| action.op == op)
678 .count();
679 assert_eq!(count, 1, "{} is {count} actions", op.name());
680 }
681 for tool in TOOLS {
682 let mut names = std::collections::HashSet::new();
683 for action in tool.actions {
684 assert!(names.insert(action.name), "{}.{} twice", tool.name, action.name);
685 }
686 if let Some(default) = tool.default_action {
687 assert!(tool.action(default).is_some(), "{}", tool.name);
688 }
689 }
Usage, Billing settings and prepaid AI credit; fixes from the UX audit690 assert!(TOOLS.len() <= 17, "{} tools", TOOLS.len());
Thirteen MCP tools and classic token scopes; agents rate their confidence and can be put on an issue in one step691 }
692
693 #[test]
694 fn every_operation_needs_exactly_one_scope_or_none() {
695 use g1t_contracts::scopes::OPERATIONS;
696 for op in Op::ALL {
697 let mapped = OPERATIONS.iter().filter(|(name, _)| *name == op.name()).count();
698 let free = NO_SCOPE.contains(&op.name());
699 assert_eq!(mapped + usize::from(free), 1, "{}", op.name());
700 }
701 for (name, _) in OPERATIONS {
702 assert!(Op::by_name(name).is_some(), "{name} is not an operation");
703 }
704 }
705
706 #[test]
707 fn each_tool_schema_is_valid_with_one_branch_per_action() {
708 for tool in TOOLS {
709 let actions: Vec<&Action> = tool.actions.iter().collect();
710 let flat = tool.input_schema(&actions);
711 assert_eq!(flat["type"], "object");
712 assert!(flat.get("oneOf").is_none(), "no oneOf at the top level");
713 let listed: Vec<&str> = flat["properties"]["action"]["enum"]
714 .as_array()
715 .unwrap()
716 .iter()
717 .map(|name| name.as_str().unwrap())
718 .collect();
719 assert_eq!(listed, tool.actions.iter().map(|action| action.name).collect::<Vec<_>>());
720 for action in tool.actions {
721 for field in action.op.required() {
722 assert!(flat["properties"].get(&field).is_some(), "{}.{}: {field}", tool.name, action.name);
723 }
724 }
725 let keyed = tool.discriminated(&actions);
726 let branches = keyed["oneOf"].as_array().unwrap();
727 assert_eq!(branches.len(), tool.actions.len());
728 for (branch, action) in branches.iter().zip(tool.actions) {
729 assert_eq!(branch["properties"]["action"]["const"], action.name);
730 for field in branch["required"].as_array().unwrap() {
731 assert!(branch["properties"].get(field.as_str().unwrap()).is_some(), "{}.{}: {field}", tool.name, action.name);
732 }
733 }
734 // A well-formed JSON Schema object throughout.
735 let text = serde_json::to_string(&flat).unwrap();
736 assert!(serde_json::from_str::<Value>(&text).is_ok());
737 }
738 }
739
740 #[test]
741 fn a_read_only_token_sees_read_actions_only() {
742 let access = token(Preset::ReadOnly.scopes());
743 let gate = Gate::Token(&access);
744 for tool in TOOLS {
745 for action in tool.visible(&gate) {
746 assert!(reads_only(action.op), "{}.{}", tool.name, action.name);
747 }
748 }
749 let tools = listed(&gate);
750 for tool in &tools {
751 assert_eq!(tool["annotations"]["readOnlyHint"], true, "{}", tool["name"]);
752 assert_eq!(tool["annotations"]["destructiveHint"], false);
753 }
754 let issue = tools.iter().find(|tool| tool["name"] == "issue").unwrap();
Teams and CODEOWNERS, labels and milestones, dependency updates, the security suite, and a clearer top bar755 assert_eq!(issue["inputSchema"]["properties"]["action"]["enum"], json!(["list", "get", "labels"]));
Thirteen MCP tools and classic token scopes; agents rate their confidence and can be put on an issue in one step756 // Nothing of the agent tool is a read.
757 assert!(!tools.iter().any(|tool| tool["name"] == "agent"));
758 }
759
760 #[test]
761 fn a_narrow_token_sees_only_its_tools() {
762 let access = token(Some(vec![Scope::IssuesWrite]));
763 let names: Vec<Value> = listed(&Gate::Token(&access)).into_iter().map(|tool| tool["name"].clone()).collect();
Teams and CODEOWNERS, labels and milestones, dependency updates, the security suite, and a clearer top bar764 // Labels and milestones are the repository's, managed with issues:write.
765 assert_eq!(names, vec![json!("repository"), json!("issue"), json!("plan"), json!("account")]);
API: notifications over REST and MCP, with notifications scopes766 // Notifications are a resource of their own: reading them lists
767 // only what reads.
768 let reader = token(Some(vec![Scope::NotificationsRead]));
769 let tools = listed(&Gate::Token(&reader));
770 let notifications = tools.iter().find(|tool| tool["name"] == "notifications").unwrap();
771 assert_eq!(
772 notifications["inputSchema"]["properties"]["action"]["enum"],
773 json!(["list", "get", "subscription", "watching", "watched"])
774 );
775 assert_eq!(notifications["annotations"]["readOnlyHint"], true);
Thirteen MCP tools and classic token scopes; agents rate their confidence and can be put on an issue in one step776 let full = token(None);
777 assert_eq!(listed(&Gate::Token(&full)).len(), TOOLS.len());
778 assert_eq!(listed(&Gate::Everything).len(), TOOLS.len());
779 }
780
781 #[test]
782 fn a_tool_that_can_destroy_says_so() {
783 let tools = listed(&Gate::Everything);
784 let repository = tools.iter().find(|tool| tool["name"] == "repository").unwrap();
785 assert_eq!(repository["annotations"]["destructiveHint"], true);
786 assert_eq!(repository["annotations"]["readOnlyHint"], false);
787 let memory = tools.iter().find(|tool| tool["name"] == "memory").unwrap();
788 assert_eq!(memory["annotations"]["destructiveHint"], false);
789 }
790
791 #[test]
792 fn calls_resolve_to_their_operation_or_say_what_is_missing() {
793 let issue = Tool::by_name("issue").unwrap();
794 assert_eq!(resolve(issue, &json!({ "action": "get", "repo": "a/b", "number": 1 })), Ok(Op::GetIssue));
795 assert_eq!(resolve(issue, &json!({ "action": "get", "repo": "a/b" })), Err("issue.get needs number.".to_owned()));
796 assert!(resolve(issue, &json!({})).unwrap_err().starts_with("Give an action"));
797 assert!(resolve(issue, &json!({ "action": "explode" })).unwrap_err().contains("no action explode"));
798 let search = Tool::by_name("search").unwrap();
799 assert_eq!(resolve(search, &json!({ "query": "x" })), Ok(Op::Search));
800 let account = Tool::by_name("account").unwrap();
801 assert_eq!(resolve(account, &json!({})), Ok(Op::Whoami));
802 }
803
Teams and CODEOWNERS, labels and milestones, dependency updates, the security suite, and a clearer top bar804 #[test]
805 fn teams_are_one_tool_and_a_workspace_reader_sees_only_its_reads() {
806 let team = Tool::by_name("team").unwrap();
807 let names: Vec<&str> = team.actions.iter().map(|action| action.name).collect();
808 assert_eq!(
809 names,
810 [
811 "list",
812 "get",
813 "create",
814 "update",
815 "delete",
816 "list_members",
817 "set_member",
818 "remove_member",
819 "list_child_teams",
820 "list_repos",
821 "set_repo",
822 "remove_repo",
823 "set_review_assignment",
824 "list_user_teams",
825 ]
826 );
827 let reader = token(Some(vec![Scope::WorkspaceRead]));
828 let tools = listed(&Gate::Token(&reader));
829 let listed_team = tools.iter().find(|tool| tool["name"] == "team").unwrap();
830 assert_eq!(
831 listed_team["inputSchema"]["properties"]["action"]["enum"],
832 json!(["list", "get", "list_members", "list_child_teams", "list_repos", "list_user_teams"])
833 );
834 assert_eq!(listed_team["annotations"]["readOnlyHint"], true);
835 // A team's role on a repository is who has access.
836 let admin = token(Some(vec![Scope::WorkspaceAdmin]));
837 let tools = listed(&Gate::Token(&admin));
838 let listed_team = tools.iter().find(|tool| tool["name"] == "team").unwrap();
839 let actions = listed_team["inputSchema"]["properties"]["action"]["enum"].as_array().unwrap();
840 assert!(actions.contains(&json!("set_review_assignment")) && !actions.contains(&json!("set_repo")));
841 let access = token(Some(vec![Scope::AccessAdmin]));
842 let tools = listed(&Gate::Token(&access));
843 let listed_team = tools.iter().find(|tool| tool["name"] == "team").unwrap();
844 assert_eq!(listed_team["inputSchema"]["properties"]["action"]["enum"], json!(["set_repo", "remove_repo"]));
845 // Both kinds of role a schema names are offered.
846 let roles = &listed(&Gate::Everything).into_iter().find(|tool| tool["name"] == "team").unwrap()["inputSchema"]
847 ["properties"]["role"]["enum"];
848 for role in ["member", "maintainer", "read", "admin"] {
849 assert!(roles.as_array().unwrap().contains(&json!(role)), "{role}");
850 }
851 assert_eq!(
852 resolve(team, &json!({ "action": "set_repo", "workspace": "acme", "team": "backend", "repo": "rocket" })),
853 Err("team.set_repo needs role.".to_owned())
854 );
855 }
856
857 #[test]
858 fn reviewers_and_code_owners_are_actions_of_their_tools() {
859 let pull = Tool::by_name("pull_request").unwrap();
860 assert_eq!(
861 resolve(pull, &json!({ "action": "request_reviewers", "repo": "a/b", "number": 1, "team_reviewers": ["backend"] })),
862 Ok(Op::RequestReviewers)
863 );
864 assert_eq!(pull.action("remove_requested_reviewers").map(|action| action.op), Some(Op::RemoveRequestedReviewers));
865 let repository = Tool::by_name("repository").unwrap();
866 assert_eq!(resolve(repository, &json!({ "action": "codeowners", "repo": "a/b" })), Ok(Op::GetCodeownersErrors));
867 assert!(reads_only(Op::GetCodeownersErrors));
868 assert!(!reads_only(Op::RequestReviewers));
869 }
870
Thirteen MCP tools and classic token scopes; agents rate their confidence and can be put on an issue in one step871 /// How much smaller `tools/list` is than one tool per operation. Run
872 /// with `--nocapture` to see the numbers.
873 #[test]
874 fn the_tool_list_is_much_smaller_than_one_tool_per_operation() {
875 let before: Vec<Value> = Op::ALL
876 .into_iter()
877 .map(|op| json!({ "name": op.name(), "description": op.description(), "inputSchema": op.input() }))
878 .collect();
879 let after = listed(&Gate::Everything);
880 let before_bytes = serde_json::to_string(&json!({ "tools": before })).unwrap().len();
881 let after_bytes = serde_json::to_string(&json!({ "tools": after })).unwrap().len();
882 let agent = token(Preset::Agent.scopes());
883 let agent_bytes = serde_json::to_string(&json!({ "tools": listed(&Gate::Token(&agent)) })).unwrap().len();
884 let read = token(Preset::ReadOnly.scopes());
885 let read_bytes = serde_json::to_string(&json!({ "tools": listed(&Gate::Token(&read)) })).unwrap().len();
886 println!(
887 "tools/list: before {} tools, {before_bytes} bytes (~{} tokens); after {} tools, {after_bytes} bytes (~{} tokens); agent preset {agent_bytes} bytes (~{} tokens); read only {read_bytes} bytes (~{} tokens)",
888 before.len(),
889 before_bytes / 4,
890 after.len(),
891 after_bytes / 4,
892 agent_bytes / 4,
893 read_bytes / 4,
894 );
895 assert!(after_bytes * 2 < before_bytes, "{after_bytes} vs {before_bytes}");
896 }
897}

This file's history is long; its oldest lines are credited to the oldest commit read.