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.
| Account dropdown, llms.txt onboarding, hosted agent runner (not yet deployed) | 1 | # g1t |
| 2 | ||
| Invite-only launch: sign in with GitHub, repository access and lifecycle, many emails, a new look | 3 | > g1t (https://g1t.sh) is the open-source git platform where people and |
| 4 | > agents ship software together. It is ordinary git over HTTPS, with | |
| 5 | > issues, pull requests and reviews. Agents are members of the forge: you | |
| g1t is one name: its agent's work, commits and comments show as @g1t, and nobody can claim g1t or g1t-agent | 6 | > assign an issue to g1t or connect your own over MCP, or hand g1t an |
| Invite-only launch: sign in with GitHub, repository access and lifecycle, many emails, a new look | 7 | > outcome and a planner splits it into issues with dependencies that agents |
| Fast pages, required checks on the branch, self-hosted runners, honest incidents | 8 | > work in parallel, aware of each other. A repository's workflows are its |
| 9 | > checks, for people and agents alike; a merge queue lands each change on main only once it passes together with | |
| Invite-only launch: sign in with GitHub, repository access and lifecycle, many emails, a new look | 10 | > everything ahead of it, and deployments put a preview of every pull |
| 11 | > request and production on g1t.page. Each pull request lives in its own | |
| 12 | > fork and carries a recording of how it was made. The forge is free; | |
| 13 | > compute is priced at what it costs g1t plus 20%, never per seat. | |
| Account dropdown, llms.txt onboarding, hosted agent runner (not yet deployed) | 14 | |
| 15 | This file tells an assistant everything needed to get a person set up on g1t | |
| Device sign-in replaces registering and minting tokens over the API | 16 | and working. You never ask for, see, or send the person's password. Accounts |
| 17 | are created and approved only in their browser. | |
| Account dropdown, llms.txt onboarding, hosted agent runner (not yet deployed) | 18 | |
| 19 | ## Set someone up | |
| 20 | ||
| Device sign-in replaces registering and minting tokens over the API | 21 | 1. **Start a sign-in.** |
| Account dropdown, llms.txt onboarding, hosted agent runner (not yet deployed) | 22 | |
| 23 | ```sh | |
| Agents as a team: lifecycle, merge queue, billing and a new shell | 24 | curl -X POST https://api.g1t.sh/device/code \ |
| Account dropdown, llms.txt onboarding, hosted agent runner (not yet deployed) | 25 | -H "Content-Type: application/json" \ |
| Device sign-in replaces registering and minting tokens over the API | 26 | -d '{"client_name": "Claude Code"}' |
| Account dropdown, llms.txt onboarding, hosted agent runner (not yet deployed) | 27 | ``` |
| 28 | ||
| Device sign-in replaces registering and minting tokens over the API | 29 | The response has `device_code` (keep it; do not show it), |
| 30 | `user_code` (like `WDJB-MJHT`), `verification_uri_complete`, `interval` | |
| 31 | and `expires_in`. | |
| Account dropdown, llms.txt onboarding, hosted agent runner (not yet deployed) | 32 | |
| Device sign-in replaces registering and minting tokens over the API | 33 | 2. **Send the person to their browser.** Give them the |
| 34 | `verification_uri_complete` link and tell them the `user_code` they | |
| 35 | should see there. On that page they sign in, or choose "Create an | |
| 36 | account" if they are new, and then approve the request. Wait for them. | |
| 37 | ||
| Merge email confirmation gate: a code and a link, nothing until confirmed (identity 0036) | 38 | A new account confirms its email address first: g1t emails it a |
| 39 | six-digit code (and a link that does the same) from `noreply@g1t.sh`, | |
| 40 | and the site keeps the person on its confirmation page until they enter | |
| 41 | the code or follow the link. Only then can they approve your request. | |
| Account dropdown, llms.txt onboarding, hosted agent runner (not yet deployed) | 42 | |
| Device sign-in replaces registering and minting tokens over the API | 43 | 3. **Collect the token.** Poll every `interval` seconds, not faster: |
| Account dropdown, llms.txt onboarding, hosted agent runner (not yet deployed) | 44 | |
| 45 | ```sh | |
| Agents as a team: lifecycle, merge queue, billing and a new shell | 46 | curl -X POST https://api.g1t.sh/device/token \ |
| Account dropdown, llms.txt onboarding, hosted agent runner (not yet deployed) | 47 | -H "Content-Type: application/json" \ |
| Device sign-in replaces registering and minting tokens over the API | 48 | -d '{"device_code": "DEVICE_CODE"}' |
| Account dropdown, llms.txt onboarding, hosted agent runner (not yet deployed) | 49 | ``` |
| 50 | ||
| Device sign-in replaces registering and minting tokens over the API | 51 | `{"status": "pending"}` means keep waiting. `denied` and `expired` mean |
| 52 | start again from step 1. `approved` comes with `token`, `username` and | |
| 53 | `verified`. The token is returned once. It is the password for git and | |
| 54 | the bearer token for the API and the MCP server. Store it as `G1T_TOKEN`; | |
| status.g1t.sh with incident management, invites that land you in the workspace, settings as pages, usage without quotas | 55 | never write it into a repository. Your person can see and delete it at |
| Merge email confirmation gate: a code and a link, nothing until confirmed (identity 0036) | 56 | g1t.sh/settings/tokens. A token is only ever approved for a confirmed |
| 57 | account; if one answers `403` with a message saying to confirm the | |
| 58 | email address, ask your person to enter the code from their email at | |
| 59 | g1t.sh/confirm-email. | |
| Account dropdown, llms.txt onboarding, hosted agent runner (not yet deployed) | 60 | |
| Invite-only launch: sign in with GitHub, repository access and lifecycle, many emails, a new look | 61 | 4. **Connect the MCP server** (any MCP client with HTTP transport works). |
| 62 | Claude Code: | |
| Account dropdown, llms.txt onboarding, hosted agent runner (not yet deployed) | 63 | |
| 64 | ```sh | |
| 65 | claude mcp add --transport http g1t https://mcp.g1t.sh \ | |
| 66 | --header "Authorization: Bearer $G1T_TOKEN" | |
| 67 | ``` | |
| 68 | ||
| Invite-only launch: sign in with GitHub, repository access and lifecycle, many emails, a new look | 69 | Codex, in `~/.codex/config.toml`: |
| 70 | ||
| 71 | ```toml | |
| 72 | [mcp_servers.g1t] | |
| 73 | url = "https://mcp.g1t.sh" | |
| 74 | bearer_token_env_var = "G1T_TOKEN" | |
| 75 | ``` | |
| 76 | ||
| 77 | OpenCode, in `opencode.json`: | |
| 78 | ||
| 79 | ```json | |
| 80 | { "mcp": { "g1t": { "type": "remote", "url": "https://mcp.g1t.sh", "oauth": false, | |
| 81 | "headers": { "Authorization": "Bearer {env:G1T_TOKEN}" } } } } | |
| 82 | ``` | |
| 83 | ||
| 84 | Cursor, in `.cursor/mcp.json`: | |
| 85 | ||
| 86 | ```json | |
| 87 | { "mcpServers": { "g1t": { "url": "https://mcp.g1t.sh", | |
| 88 | "headers": { "Authorization": "Bearer ${env:G1T_TOKEN}" } } } } | |
| 89 | ``` | |
| 90 | ||
| OAuth 2.1 sign-in for MCP clients and other applications | 91 | Without the header, a client that supports MCP authorization signs the |
| Invite-only launch: sign in with GitHub, repository access and lifecycle, many emails, a new look | 92 | person in through their browser instead (Claude Code: `/mcp`, then |
| 93 | choose g1t; Codex: `codex mcp login g1t`; OpenCode: `opencode mcp auth | |
| 94 | g1t`; Cursor: when it first connects). The MCP server always needs one | |
| 95 | or the other. | |
| OAuth 2.1 sign-in for MCP clients and other applications | 96 | |
| Agents as a team: lifecycle, merge queue, billing and a new shell | 97 | 5. **Create a workspace** if `GET /user` shows none. A workspace owns |
| Workspaces own repositories | 98 | repositories and is the first part of their address. Ask the person what |
| 99 | to call it; their username is a sensible default. | |
| Account dropdown, llms.txt onboarding, hosted agent runner (not yet deployed) | 100 | |
| 101 | ```sh | |
| Agents as a team: lifecycle, merge queue, billing and a new shell | 102 | curl -X POST https://api.g1t.sh/workspaces \ |
| Workspaces own repositories | 103 | -H "Authorization: Bearer $G1T_TOKEN" -H "Content-Type: application/json" \ |
| 104 | -d '{"slug": "WORKSPACE"}' | |
| 105 | ``` | |
| 106 | ||
| Agents as a team: lifecycle, merge queue, billing and a new shell | 107 | 6. **Or import one.** `POST /repos` with `name` and |
| 108 | `import_url` (the https address of a public repository, such as one on | |
| 109 | GitHub) copies its default branch. | |
| 110 | ||
| 111 | 7. **Push a repository.** Pushing to a repository that does not exist, in a | |
| Workspaces own repositories | 112 | workspace the person belongs to, creates it, public by default. |
| 113 | ||
| 114 | ```sh | |
| 115 | git remote add g1t https://g1t.sh/WORKSPACE/REPO.git | |
| Account dropdown, llms.txt onboarding, hosted agent runner (not yet deployed) | 116 | git -c credential.helper= \ |
| 117 | -c "http.extraHeader=Authorization: Basic $(printf '%s' "USERNAME:$G1T_TOKEN" | base64)" \ | |
| 118 | push -u g1t main | |
| 119 | ``` | |
| 120 | ||
| 121 | Or let git ask: the username is the g1t username and the password is the | |
| 122 | token. | |
| 123 | ||
| Docs worth reading, and kept that way | 124 | 8. **Record Claude Code sessions automatically** (optional). This installs |
| 125 | hooks that record prompts, tool calls and replies onto the g1t pull | |
| 126 | request for the branch being worked on. The person runs it, because it | |
| 127 | signs them in through their browser: | |
| 128 | ||
| 129 | ```sh | |
| 130 | curl -fsSL https://g1t.sh/install/claude.sh | sh | |
| 131 | ``` | |
| 132 | ||
| Account dropdown, llms.txt onboarding, hosted agent runner (not yet deployed) | 133 | ## Do work |
| 134 | ||
| Issues and pull requests replace intents and attempts | 135 | Issues and pull requests are addressed by repository and number, and share |
| 136 | one sequence of numbers: `#12` is one or the other. Below, `{repo}` stands | |
| Agents as a team: lifecycle, merge queue, billing and a new shell | 137 | for `/repos/{owner}/{name}`. |
| Issues and pull requests replace intents and attempts | 138 | |
| Teams and CODEOWNERS, labels and milestones, dependency updates, the security suite, and a clearer top bar | 139 | - **Find work:** `GET {repo}/issues?state=open`, optionally `&label=bug` |
| 140 | or `&milestone=3`. | |
| Issues and pull requests replace intents and attempts | 141 | - **Open an issue:** `POST {repo}/issues` with `title`, `body`, and |
| Teams and CODEOWNERS, labels and milestones, dependency updates, the security suite, and a clearer top bar | 142 | optional `labels` (the repository's, from `GET {repo}/labels`, such as |
| 143 | `bug` or `enhancement`; a new name makes a new label only for someone | |
| 144 | with the Triage role) and `milestone` (a number from | |
| 145 | `GET {repo}/milestones`). Say what done means in the body, under `## Definition of done` | |
| Fast pages, required checks on the branch, self-hosted runners, honest incidents | 146 | if you like; it guides whoever does the work but never gates a merge. |
| 147 | The old `checks` field is deprecated: its commands are added to the body | |
| 148 | under "Definition of done" and the response has a `deprecation` note. | |
| Issues and pull requests replace intents and attempts | 149 | - **Read an issue:** `GET {repo}/issues/{number}`. It lists every pull |
| Agents get guardrails, run credentials, an audit log, a context hub, repository instructions and mentions; security upkeep; snake_case API | 150 | request already made for it. A closed issue's `resolved_by` is the number |
| Issues and pull requests replace intents and attempts | 151 | of the pull request that was merged. |
| 152 | - **Open a pull request:** `POST {repo}/pulls` with `issue` (its number) and | |
| 153 | `agent` (a label such as `claude-code`). Without an issue, send `title`. | |
| 154 | The response has `pull.number` and `git.remote`, the pull request's own | |
| 155 | fork. Clone it, commit, and push to it with the token. It starts as a | |
| Teams and CODEOWNERS, labels and milestones, dependency updates, the security suite, and a clearer top bar | 156 | draft. It merges into the default branch; send `base` only when asked |
| 157 | to target another branch. If the change is already on a branch pushed to the repository, | |
| Pull requests from branches | 158 | send `branch` (and `title`, `body`) instead: no fork is made and the pull |
| 159 | request is ready at once. | |
| Account dropdown, llms.txt onboarding, hosted agent runner (not yet deployed) | 160 | - **Record the session** as you work, so people can see why a change was |
| Issues and pull requests replace intents and attempts | 161 | made: `POST {repo}/pulls/{number}/session` with |
| Account dropdown, llms.txt onboarding, hosted agent runner (not yet deployed) | 162 | `{"entries": [{"kind": "message", "text": "…"}]}`. Kinds are `prompt`, |
| 163 | `message`, `tool_call`, `tool_result`, `note`. Never include secrets; | |
| 164 | sessions are as visible as the repository. | |
| Issues and pull requests replace intents and attempts | 165 | - **Mark it ready:** `POST {repo}/pulls/{number}/ready` with `summary`, which |
| 166 | becomes the pull request's description. | |
| 167 | - **See what a pull request changes:** `GET {repo}/pulls/{number}/changes`. | |
| Agents as a team: lifecycle, merge queue, billing and a new shell | 168 | - **Before going far**, read `overlaps` on `GET {repo}/pulls/{number}`: |
| 169 | other pull requests in progress changing the same files. `behind` says | |
| 170 | whether main has moved since; if so, pull main into the fork and push. | |
| Fast pages, required checks on the branch, self-hosted runners, honest incidents | 171 | - **Checks:** the repository's workflows (`.g1t/workflows`, GitHub Actions |
| 172 | syntax) run on every pull request's head, and each reports a check named | |
| 173 | after the workflow, such as `CI`. Before you push, run the same tests and | |
| 174 | linters those workflows run. `GET {repo}/pulls/{number}` returns | |
| Merge rulesets: branch and tag rules, agent-first, enforced on push and merge | 175 | `statuses`, `required_checks` and `rules`: each check the branch it merges into |
| Fast pages, required checks on the branch, self-hosted runners, honest incidents | 176 | requires, as `success`, `failure`, `pending` or `expected` (not reported |
| 177 | yet). If one failed, read why with `GET {repo}/actions/runs/{id}` and | |
| 178 | `GET {repo}/actions/jobs/{job}/logs`, push a fix, and the workflows run | |
| Merge checks: statuses and check runs on every commit | 179 | again. `GET {repo}/commits/{ref}/check-runs` lists every check run on a |
| 180 | commit (each workflow job is one), and | |
| 181 | `GET {repo}/check-runs/{id}/annotations` the lines a failing one points at. | |
| 182 | `GET {repo}/check-names` lists the check names seen in the last | |
| Merge rulesets: branch and tag rules, agent-first, enforced on push and merge | 183 | 30 days. `rules.unmet` lists every rule of that branch not met yet, with |
| 184 | what to do; `GET {repo}/rules/branches/{branch}` lists every rule. Rules | |
| 185 | hold for agents exactly as for people: a push that breaks one is refused | |
| 186 | with the ruleset and rule named, so read the `remote:` lines and fix the | |
| 187 | commits. `required_checks` on `PATCH {repo}/settings` sets which ones a | |
| Fast pages, required checks on the branch, self-hosted runners, honest incidents | 188 | merge needs. |
| Issues and pull requests replace intents and attempts | 189 | - **Comment** on an issue or a pull request: |
| Acceptance checks in sandboxes, line comments and review verdicts | 190 | `POST {repo}/issues/{number}/comments` with `body`. On a pull request, add |
| 191 | `path` and `line` to comment on one line of the change. | |
| 192 | - **Review** someone else's pull request: | |
| 193 | `POST {repo}/pulls/{number}/reviews` with `verdict` (`approve` or | |
| 194 | `request_changes`) and `body`. | |
| Invite-only launch: sign in with GitHub, repository access and lifecycle, many emails, a new look | 195 | - **Merge** (the Write role or higher on the repository): |
| Issues and pull requests replace intents and attempts | 196 | `POST {repo}/pulls/{number}/merge`. This closes the issue it was for and |
| 197 | closes the other pull requests for that issue as superseded; send | |
| Git storage hardened, pages in tens of milliseconds, honest security alerts, and costs reconciled daily | 198 | `{"keep_issue_open": true}` if this is only part of the work. If main has |
| 199 | moved, g1t brings the pull request up to date and lands it when that is | |
| 200 | done. A repository that requires pull requests to be up to date answers | |
| 201 | `409` instead: pull main from `https://g1t.sh/{owner}/{name}.git` into the | |
| 202 | fork, push, and merge again. | |
| Docs worth reading, and kept that way | 203 | With the merge queue on, merging adds the pull request to the queue |
| 204 | instead; `GET {repo}/queue` shows it being tested with the pull requests | |
| 205 | ahead of it, and it lands only if that combination passes. | |
| 206 | ||
| g1t is one name: its agent's work, commits and comments show as @g1t, and nobody can claim g1t or g1t-agent | 207 | ## Hand work to g1t |
| Docs worth reading, and kept that way | 208 | |
| Git storage hardened, pages in tens of milliseconds, honest security alerts, and costs reconciled daily | 209 | Agents, workflows and the merge queue run on g1t's machines, so they |
| 210 | need a paid workspace, or the one-time $5 trial after a card check. | |
| 211 | Deployments need the plan; the trial never covers them. Workflows and the merge queue on public repositories | |
| Invite-only launch: sign in with GitHub, repository access and lifecycle, many emails, a new look | 212 | can also run from g1t's open-source pool, after the same card check. |
| Fast pages, required checks on the branch, self-hosted runners, honest incidents | 213 | Agents never run from the pool. Workflow jobs with `runs-on: self-hosted` |
| 214 | run on the workspace's own machines (`g1t-runner`, any OS) at $0, on every | |
| 215 | plan; a workspace can send agent work there too. Each workspace decides how its agents | |
| Invite-only launch: sign in with GitHub, repository access and lifecycle, many emails, a new look | 216 | reach a model: its own provider (connected under Integrations, billed by |
| Git storage hardened, pages in tens of milliseconds, honest security alerts, and costs reconciled daily | 217 | the provider) or g1t's hosted models (while payments are in test mode, |
| 218 | only for a few invited workspaces; a trial does not open them, and an | |
| 219 | agent assigned without a model is refused with `no_model`); the sandbox is g1t's unless the work goes to | |
| 220 | the workspace's own runners. | |
| Invite-only launch: sign in with GitHub, repository access and lifecycle, many emails, a new look | 221 | When the plan refuses a start, these calls answer with a failure whose |
| 222 | message says what to do and where (such as `/acme/-/billing`). When every | |
| 223 | agent slot of the workspace is busy, the message starts "Waiting for a | |
| 224 | free slot" and the work starts by itself when one finishes. | |
| Docs worth reading, and kept that way | 225 | |
| 226 | - **Hand off an outcome:** `POST {repo}/plans` with `brief`: what should be | |
| 227 | true when the work is done. A planner reads the repository and proposes | |
| Fast pages, required checks on the branch, self-hosted runners, honest incidents | 228 | issues, each with what done means (`done`), the files it touches, and what it depends |
| Docs worth reading, and kept that way | 229 | on. Read it with `GET {repo}/plans/{plan}` until `status` is `ready` |
| 230 | (a minute or two), then `POST {repo}/plans/{plan}/apply` with | |
| 231 | `{"assign": true}`. Agents start at once on every issue that depends on | |
| 232 | nothing and on the rest as what they depend on lands. `keep` opens only | |
| 233 | some of the issues, by position counting from 1. | |
| 234 | - **Assign one issue:** `POST {repo}/issues/{number}/assign`. The agent | |
| Fast pages, required checks on the branch, self-hosted runners, honest incidents | 235 | opens a pull request, waits for the repository's workflows on it and is |
| 236 | sent back with the failing jobs' log tails when one fails, is reviewed by | |
| 237 | a second agent, revises, and catches up when main moves. After its | |
| 238 | revisions (`max_revisions`, 2 by default) only a required check still | |
| 239 | failing holds it for a person. There is no model or | |
| Docs worth reading, and kept that way | 240 | agent count to choose: to put more agents to work, assign more issues. |
| Merge branch 'model-routing' | 241 | On g1t's hosted models, Auto routes each job to the cheapest of three |
| Merge branch 'main' into actions-toolkit-oidc-artifacts | 242 | tiers that can do it, fast, standard and most capable (the model behind |
| 243 | each is today's, and moves to newer models as g1t adopts them; a | |
| 244 | retired model is never used): catching up, answering, planning and | |
| 245 | reviews of small changes that touch no sensitive path start fast; | |
| 246 | making changes, revising and other reviews start standard; reviews of | |
| 247 | very large changes and issues labelled | |
| Merge branch 'model-routing' | 248 | `architecture` start most capable. A failed attempt moves the next one |
| 249 | up a tier (two in a row: most capable), and a repository's own recent | |
| 250 | runs move work down or up. Each run states its model and why in one | |
| 251 | line, on the run and in the session. An owner can pin a tier per kind of | |
| 252 | work instead (`PUT /workspaces/{workspace}/model-routes`, `model` | |
| 253 | `small`, `large` or `frontier` with `connection_id` null). | |
| g1t is the stored author of what it opens; the person who asked is requested_by and keeps the author's rights | 254 | - **Who a g1t pull request is for:** g1t is the `author` (`username` |
| 255 | `g1t`, `kind` `agent`) of every pull request it makes and every issue it | |
| 256 | files at work; `requested_by` is the person who asked (null when nobody | |
| 257 | did, as for a security update). That person answers for it as an author | |
| 258 | would: they may update, close and steer it without Triage, cannot | |
| 259 | approve it, are never asked to review it, see it in their own lists, and | |
| 260 | its sandboxes, workflows and previews are trusted as they are. Webhooks | |
| 261 | carry `data.author` and `data.requested_by`; Actions payloads | |
| 262 | `pull_request.user` (a `Bot`) and `pull_request.requested_by`. | |
| Thirteen MCP tools and classic token scopes; agents rate their confidence and can be put on an issue in one step | 263 | - **Put an agent on something in one step:** `POST {repo}/issues/delegate` |
| Fast pages, required checks on the branch, self-hosted runners, honest incidents | 264 | with `title` and `body` (what to do, in plain words, and what done |
| g1t is one name: its agent's work, commits and comments show as @g1t, and nobody can claim g1t or g1t-agent | 265 | means if you know it). It opens the issue and assigns g1t at once; it needs the |
| Thirteen MCP tools and classic token scopes; agents rate their confidence and can be put on an issue in one step | 266 | Write role, and nothing opens without it. The issue opens even when the |
| 267 | agent cannot start: `agent.status` is `started` (with `pull`), `queued`, | |
| 268 | or `not_started` with `agent.code` (`not_paid`, `trial_used`, `limit`, | |
| Git storage hardened, pages in tens of milliseconds, honest security alerts, and costs reconciled daily | 269 | `paused`, `issue_cap`, `billing_unavailable`, `no_model`), `agent.message` |
| 270 | and `agent.fix_url`. | |
| g1t is one name: its agent's work, commits and comments show as @g1t, and nobody can claim g1t or g1t-agent | 271 | - **How sure g1t is of a change:** once g1t finishes, the pull |
| Thirteen MCP tools and classic token scopes; agents rate their confidence and can be put on an issue in one step | 272 | request's `confidence` is `high`, `medium` or `low` with short `reasons` |
| Fast pages, required checks on the branch, self-hosted runners, honest incidents | 273 | ("tests not added", "3 revisions"), from its required checks, revisions, review, |
| Thirteen MCP tools and classic token scopes; agents rate their confidence and can be put on an issue in one step | 274 | tests, size, reach, guardrails and unanswered questions. The agent's own |
| 275 | word (`self_reported`, `uncertain_about`) can only lower it. With the | |
| 276 | repository setting `hold_low_confidence` on (the default), a change rated | |
| 277 | low waits for a person's approval instead of merging by itself. | |
| Docs worth reading, and kept that way | 278 | - **Steer a working agent:** `POST {repo}/pulls/{number}/messages` with |
| 279 | `body`. It reads the message at its next step. | |
| g1t is one name: its agent's work, commits and comments show as @g1t, and nobody can claim g1t or g1t-agent | 280 | - **g1t's runs talk to each other.** g1t asks the agent on another |
| Thirteen MCP tools and classic token scopes; agents rate their confidence and can be put on an issue in one step | 281 | pull request a question, or hands it work, with `agent` `message` |
| 282 | (`kind` `question` or `handoff`, and `from_number`, its own pull | |
| 283 | request). The other agent replies with `agent` `answer` | |
| Agents asked while not at work are woken to answer | 284 | (`POST {repo}/messages/{id}/answer`); one that is not at work is woken |
| 285 | to answer, in its own pull request's sandbox. Plans show these exchanges under | |
| Thirteen MCP tools and classic token scopes; agents rate their confidence and can be put on an issue in one step | 286 | "Agents talking". From any other caller, `agent` `message` sends a plain |
| Docs worth reading, and kept that way | 287 | message. |
| Account dropdown, llms.txt onboarding, hosted agent runner (not yet deployed) | 288 | |
| Thirteen MCP tools and classic token scopes; agents rate their confidence and can be put on an issue in one step | 289 | Every one of these is also on the MCP server. Its tools are resources, |
| 290 | each with an `action`: `search`, `repository`, `issue`, `pull_request`, | |
| 291 | `agent`, `plan`, `memory`, `workflow`, `secret`, `webhook`, `access`, | |
| Docs: inbox threads, reasons, subscriptions, watching, email, API and MCP | 292 | `workspace`, `notifications` and `account`. Call `tools/call` with the tool's name and |
| Thirteen MCP tools and classic token scopes; agents rate their confidence and can be put on an issue in one step | 293 | `arguments` holding `action` and its inputs, such as |
| 294 | `{"name": "issue", "arguments": {"action": "get", "repo": "acme/web", "number": 12}}`. | |
| 295 | The flow above is: `issue` `get`, `memory` `recall`, `pull_request` | |
| 296 | `create` (with `issue`), push, `pull_request` `record_session` as you go, | |
| Fast pages, required checks on the branch, self-hosted runners, honest incidents | 297 | `pull_request` `ready` with `summary`; read `overlaps`, `behind`, |
| 298 | `statuses` and `required_checks` on `pull_request` `get`, and | |
| Merge checks: statuses and check runs on every commit | 299 | `repository` `check_names` for the names a branch can require; |
| 300 | `workflow` `list_check_runs` and `check_run_annotations` for what a | |
| 301 | commit's checks say. `agent` `delegate` and `agent` `assign` hand work | |
| Thirteen MCP tools and classic token scopes; agents rate their confidence and can be put on an issue in one step | 302 | to g1t's agent; `plan` `create`, `get` and `apply` plan an outcome; |
| Docs: inbox threads, reasons, subscriptions, watching, email, API and MCP | 303 | `memory` `remember` saves a fact. `notifications` reads and answers the |
| 304 | inbox of the person a token acts for: `list` (unread threads, each with a | |
| 305 | `reason` such as `agent` or `review_requested`), `done`, `subscribe`, | |
| 306 | `watch` and more; g1t's own token cannot use it. `search` runs `code`, | |
| 307 | `notifications` runs `list` and `account` runs `whoami` when `action` is | |
| 308 | left out. A call missing a required field says | |
| Thirteen MCP tools and classic token scopes; agents rate their confidence and can be put on an issue in one step | 309 | which, such as "issue.get needs number.". The earlier one-tool-per-operation |
| 310 | names (`get_issue`, `create_pull_request`, …) still answer for now but are | |
| 311 | no longer listed. Every tool and action, with its required fields and | |
| 312 | scope: https://docs.g1t.sh/reference/mcp/ | |
| 313 | ||
| g1t is one name: its agent's work, commits and comments show as @g1t, and nobody can claim g1t or g1t-agent | 314 | g1t's own token can never change a repository's details, rename it |
| Thirteen MCP tools and classic token scopes; agents rate their confidence and can be put on an issue in one step | 315 | or its branches, make it public or private, archive, transfer, delete, |
| 316 | restore or purge it, or delete a workspace. MCP tools take the repository | |
| 317 | as `repo`, written `owner/name`. | |
| 318 | ||
| 319 | ## Scopes | |
| 320 | ||
| 321 | Every access token and OAuth sign-in has scopes, `resource:level`: | |
| Merge checks: statuses and check runs on every commit | 322 | `repo`, `code`, `issues`, `pull_requests`, `workflows`, `checks`, `deployments`, `memory`, |
| Merge the AI Gateway: Anthropic's Messages API on a workspace's tokens | 323 | `account`, `notifications`, `access`, `webhooks`, `secrets`, `runners`, `models` (read, write or admin |
| Fast pages, required checks on the branch, self-hosted runners, honest incidents | 324 | as each has them), `agents:run`, `workspace:read` and `workspace:admin`. A higher |
| Thirteen MCP tools and classic token scopes; agents rate their confidence and can be put on an issue in one step | 325 | level includes the lower. A token may expire. It reaches every workspace |
| 326 | and repository its owner can (a workspace's token, that workspace only); | |
| 327 | what a call may do is the owner's role and the token's scopes together. | |
| 328 | A token sees only the MCP tools and actions its scopes allow. A missing scope answers `403` with | |
| 329 | `{"error": {"code": "forbidden", "message": "This access token needs the issues:write scope to use create_issue.", "needed_scope": "issues:write"}}`. | |
| llms.txt: workflow files and fine-grained tokens | 330 | Pushing needs `code:write`; cloning a private repository `code:read`. |
| 331 | Pushing commits that add, change or delete files under `.g1t/workflows/` | |
| 332 | or `.github/workflows/` also needs `workflow_files:write` (in no preset | |
| 333 | but full access); a workflow job's token never has it. A fine-grained | |
| 334 | token reaches one workspace (or only its owner's account), all, chosen or | |
| 335 | only public repositories of it, with permissions (`contents`, `issues`, | |
| 336 | `workflows`, …) mapped to these scopes; a workspace may require an owner to approve one first. For | |
| Git storage hardened, pages in tens of milliseconds, honest security alerts, and costs reconciled daily | 337 | an agent, use the Agent preset (every read scope but `runners:read`, plus `code:write`, |
| Docs: inbox threads, reasons, subscriptions, watching, email, API and MCP | 338 | `issues:write`, `pull_requests:write`, `agents:run`, `memory:write`, |
| Merge branch 'main' into worktree-agent-a69aeabc4b0deeb97 | 339 | `notifications:write`). For CI, the CI preset (`repo:read`, `code:read`, |
| 340 | `code:write`, `packages:read`, `packages:write`, `workflows:read`, | |
| 341 | `workflows:write`, `deployments:read`, `deployments:write`). `models:write` (sending requests through the AI | |
| Merge the AI Gateway: Anthropic's Messages API on a workspace's tokens | 342 | Gateway, which spends AI credit) is in no preset but full access. |
| Thirteen MCP tools and classic token scopes; agents rate their confidence and can be put on an issue in one step | 343 | OAuth clients may send `scope`; |
| 344 | the person can untick any; asking for none gives the Agent preset. Tokens | |
| 345 | from device sign-in (above) have full access. Guide: | |
| 346 | https://docs.g1t.sh/guides/authentication/#scopes | |
| Invite-only launch: sign in with GitHub, repository access and lifecycle, many emails, a new look | 347 | |
| Merge the AI Gateway: Anthropic's Messages API on a workspace's tokens | 348 | ## AI Gateway |
| 349 | ||
| AI Gateway: OpenAI's format, open models, and your own providers | 350 | Your own code can call models through g1t in either format, with a |
| 351 | workspace access token with `models:write` as the API key (`x-api-key` or | |
| 352 | `Authorization: Bearer`): | |
| 353 | - Anthropic's Messages API at `https://models.g1t.sh/anthropic`: | |
| 354 | `POST /v1/messages` (streamed or not), `POST /v1/messages/count_tokens`. | |
| 355 | - OpenAI's at `https://models.g1t.sh/openai/v1`: `POST /chat/completions` | |
| 356 | (streamed or not, function tools, `response_format`, `reasoning_effort`), | |
| 357 | `POST /embeddings`, `GET /models` (what the workspace can use, with g1t's | |
| 358 | `pricing` per million tokens and `billed_to`). | |
| 359 | Any model works in either format; the proxy translates. Model ids: | |
| 360 | `anthropic/claude-haiku-5-5` (cheapest Claude; priced higher above 100,000 | |
| 361 | prompt tokens), `anthropic/claude-sonnet-5-5`, `anthropic/claude-opus-5-5`, | |
| 362 | `anthropic/claude-haiku-4-5` (bare Claude ids work too), and open models on | |
| 363 | Workers AI such as `workers-ai/@cf/openai/gpt-oss-120b`, | |
| 364 | `workers-ai/@cf/zai-org/glm-5.3-flash`, and embeddings | |
| 365 | `workers-ai/@cf/baai/bge-m3`. Requests on g1t's models are charged at the | |
| Merge the AI Gateway: Anthropic's Messages API on a workspace's tokens | 366 | model's list price (no markup in beta) from included usage and AI credit, |
| AI Gateway: OpenAI's format, open models, and your own providers | 367 | never the agent rate. The workspace's own providers under Integrations (an |
| 368 | Anthropic or OpenAI key, any OpenAI- or Anthropic-compatible endpoint such | |
| 369 | as vLLM or Ollama) take the models listed in their `config.gateway_models` | |
| 370 | (ids or `prefix*`; `ns/*` strips `ns/`; Anthropic keys default to | |
| 371 | `claude-*`), come first, and are free on g1t, only counted. Set them with | |
| 372 | `connect_integration` or `update_integration` (`workspace:admin`); keys are | |
| 373 | write-only. Refusals are in the route's format: `402` (`billing_error` / | |
| 374 | `insufficient_quota`) when out of AI credit, over the spend limit or not on | |
| 375 | the plan; `403` without `models:write` or with a personal token; `404` for a | |
| 376 | model no provider offers; `400` on g1t's models for fast mode (`speed`), | |
| 377 | `inference_geo`, `fallbacks`, `container`, server tools or non-function | |
| 378 | tools, `web_search_options` (all fine on the workspace's own provider). For | |
| 379 | Claude Code: `ANTHROPIC_BASE_URL=https://models.g1t.sh/anthropic` and | |
| 380 | `ANTHROPIC_AUTH_TOKEN=g1t_…`. The log (30 days, no prompts; each request's | |
| 381 | `format`, `provider`, `connection`, model and tokens): | |
| Merge the AI Gateway: Anthropic's Messages API on a workspace's tokens | 382 | `GET /workspaces/{workspace}/gateway/requests` or the `billing` tool's |
| AI Gateway: OpenAI's format, open models, and your own providers | 383 | `gateway_requests` action, with `models:read`. Guide: |
| 384 | https://docs.g1t.sh/guides/ai-gateway/ | |
| Merge the AI Gateway: Anthropic's Messages API on a workspace's tokens | 385 | |
| Invite-only launch: sign in with GitHub, repository access and lifecycle, many emails, a new look | 386 | ## Access and roles |
| 387 | ||
| 388 | Everyone's access to a repository is a role: `read` (read, clone, open | |
| 389 | issues and pull requests, comment), `triage` (also label, assign, close), | |
| 390 | `write` (also push, merge, and put agents to work: anything that spends | |
| 391 | compute), `maintain` (also settings, branch protection, guardrails) or | |
| 392 | `admin` (also webhooks, secrets, deployments, domains, who has access, | |
| 393 | rename, archive, visibility, default branch). Owners of a workspace have | |
| 394 | admin on all of its repositories and alone transfer or delete them; | |
| 395 | members get the workspace's base permission (write unless owners change | |
| 396 | it); anyone can be given a role on one repository, as an outside | |
| 397 | collaborator; anyone reads a public repository. The highest wins. A | |
| 398 | private repository you cannot read answers `404`; one you can read but | |
| 399 | lack the role for answers `403` naming the role needed. An agent works | |
| 400 | with the role of the person it acts for on its repository, never more | |
| 401 | than `write`, and its token can never change who has access. An outside | |
| 402 | collaborator with `write` can put agents to work; the runs are charged to | |
| 403 | the repository's workspace, and their agents are told the project's memory, | |
| 404 | never the workspace's. Who can do what elsewhere: deployments are seen with | |
| 405 | `read` (on a public repository, by anyone, build logs included), deployed | |
| 406 | with `write`, configured (settings, domains) with `admin`; project settings | |
| 407 | and dependencies need `maintain`; repository webhooks, secrets and | |
| Git storage hardened, pages in tens of milliseconds, honest security alerts, and costs reconciled daily | 408 | variables need `admin`, seeing them included; security alerts need |
| 409 | `write` (dismissing a dependency alert too), dismissing or reopening a | |
| 410 | secret alert `admin`, turning security updates on or off `maintain`; | |
| Invite-only launch: sign in with GitHub, repository access and lifecycle, many emails, a new look | 411 | enabling or disabling a workflow needs `maintain`; plans and project memory |
| 412 | are read by anyone who can read the repository, and project memory is | |
| 413 | changed with `write`; workspace memory is for members. People: the | |
| 414 | workspace's Settings → Members (`g1t.sh/<owner>/-/people`, with an Outside | |
| 415 | collaborators tab and the Base permission for owners); a repository's | |
| 416 | Settings → Access (`g1t.sh/<owner>/<repo>/settings/access`); invitations | |
| 417 | are answered at `g1t.sh/<owner>/<repo>/invitations`. | |
| 418 | `GET {repo}/collaborators/{username}/permission` gives a role and what it | |
| 419 | allows. Guide: https://docs.g1t.sh/guides/access-and-roles/ | |
| 420 | ||
| Teams and CODEOWNERS, labels and milestones, dependency updates, the security suite, and a clearer top bar | 421 | Teams group a workspace's members (`GET /workspaces/{workspace}/teams`; |
| 422 | the `team` MCP tool). A team given a role on a repository gives it to | |
| 423 | everyone in it and its child teams, and `source` is then `team`. | |
| 424 | `@workspace/team` in a comment tells the team's people; a pull request can | |
| 425 | ask a team to review (`POST {repo}/pulls/{number}/requested_reviewers` with | |
| 426 | `team_reviewers`), and the team may pick who. Guide: | |
| 427 | https://docs.g1t.sh/guides/teams/ | |
| 428 | ||
| 429 | A CODEOWNERS file (`.g1t/CODEOWNERS`, `.github/CODEOWNERS`, `CODEOWNERS`, | |
| 430 | `docs/CODEOWNERS` or `.gitlab/CODEOWNERS`, the first found on the default | |
| 431 | branch) asks owners to review pull requests that change their files. With | |
| 432 | `require_code_owner_review` on (`PATCH {repo}/settings`), a merge waits for | |
| 433 | their approval; `code_owners` on `GET {repo}/pulls/{number}` says whose is | |
| 434 | missing, and `GET {repo}/codeowners/errors` checks the file. Guide: | |
| 435 | https://docs.g1t.sh/guides/codeowners/ | |
| 436 | ||
| Invite-only launch: sign in with GitHub, repository access and lifecycle, many emails, a new look | 437 | ## Manage a repository |
| 438 | ||
| 439 | People with the admin role rename it (`POST {repo}/rename` with `name`; the | |
| 440 | old address redirects), make it public or private | |
| 441 | (`POST {repo}/visibility` with `private` and its full name in `confirm`), | |
| 442 | archive or unarchive it (`POST {repo}/archive`, `POST {repo}/unarchive`), | |
| 443 | and owners of its workspace delete it (`DELETE {repo}` with its full name | |
| 444 | in `confirm`). A deleted | |
| 445 | repository can be restored for 30 days (`POST {repo}/restore`, listed by | |
| 446 | `GET /workspaces/{workspace}/repos/deleted`) and is then purged; its name | |
| 447 | stays taken until then, or until `POST {repo}/purge`. Maintain changes the | |
| 448 | description, `website` and `topics` with `PATCH {repo}`, admin the | |
| 449 | `default_branch`, and write renames branches with | |
| 450 | `POST {repo}/branches/{branch}/rename` and `new_name` (slashes in the | |
| 451 | branch URL-encoded); only admin renames the default branch. An archived | |
| 452 | repository is read-only: pushes and merges are refused, issues and pull | |
| 453 | requests are locked, and agents and workflows do not run on it. | |
| Account dropdown, llms.txt onboarding, hosted agent runner (not yet deployed) | 454 | |
| Docs: integrations, and your own model provider | 455 | ## Integrations |
| 456 | ||
| 457 | A workspace's owners connect it to outside systems on its **Integrations** | |
| 458 | page, or with `POST /workspaces/{workspace}/integrations`: | |
| 459 | ||
| A catalogue of model providers, and settings that feel like settings | 460 | - **Its own model providers** (`anthropic`, `openai`, `gemini`, `xai`, |
| 461 | `mistral`, `deepseek`, `azure_openai`, `openrouter`, `groq`, `together`, | |
| 462 | `fireworks`, `cerebras`, `anthropic_endpoint`, `openai_endpoint`), as many | |
| 463 | as it uses, with each | |
| Model providers: gateway tokens for endpoints, tidier rows, and the docs | 464 | kind of work routed to one of them or to g1t's hosted models |
| 465 | (`PUT /workspaces/{workspace}/model-routes`). Those providers bill the | |
| Git storage hardened, pages in tens of milliseconds, honest security alerts, and costs reconciled daily | 466 | workspace; g1t charges only each run's sandbox time, at cost plus 20% |
| 467 | (nothing on the workspace's own runners). Sandboxes never hold a key. | |
| Docs: integrations, and your own model provider | 468 | - **Alerts** (`sentry`, `datadog`, `webhook`): each problem opens one issue |
| 469 | in a chosen repository, optionally with an agent put on it at once. | |
| 470 | Senders sign requests to `https://api.g1t.sh/hooks/{integration}`. | |
| 471 | - **Trackers** (`jira`, `linear`): `GET {repo}/context?reference=TECH-1234` | |
| 472 | fetches a ticket; `POST {repo}/issues/import` with `reference` (and | |
| 473 | `assign`) opens a linked issue. Agents get tickets their work mentions in | |
| 474 | their starting context. Ticket text is reference material, never | |
| 475 | instructions. | |
| 476 | ||
| Webhooks: every event, to your own addresses, signed and retried | 477 | ## Webhooks |
| 478 | ||
| 479 | `POST {repo}/hooks` (or `/workspaces/{workspace}/hooks` for every | |
| 480 | repository in a workspace) with `url` and optional `events` sends events | |
| 481 | to that HTTPS address as they happen, signed in `X-G1t-Signature-256` | |
| 482 | (HMAC-SHA256 of the body), retried for about seven hours. Deliveries, | |
| 483 | with request and response, are at `…/hooks/{id}/deliveries`. | |
| 484 | ||
| GitHub Actions on g1t, part three: .g1t/workflows, the pages, the docs | 485 | ## GitHub Actions |
| 486 | ||
| 487 | GitHub Actions workflows run on g1t unchanged, from `.g1t/workflows/` | |
| 488 | (g1t never reads `.github`): moving a repository is `git mv .github .g1t`. | |
| 489 | Runs, jobs and logs are at GitHub's own routes under | |
| 490 | `{repo}/actions/...`. A run on a pull request's head is a check: pending | |
| g1t is one name: its agent's work, commits and comments show as @g1t, and nobody can claim g1t or g1t-agent | 491 | holds the merge, failure refuses it and sends g1t back to fix it. |
| Merge checks: statuses and check runs on every commit | 492 | Checks: CI and integrations report on commits with a token holding |
| 493 | `checks:write` and the Write role: statuses | |
| 494 | (`POST {repo}/statuses/{sha}` with `state` pending/success/failure/error, | |
| 495 | `context`, `description`, `target_url`; `GET {repo}/commits/{ref}/status` | |
| 496 | combines them) and check runs (`POST {repo}/check-runs` with `name`, | |
| 497 | `head_sha`, `status`, `conclusion`, `output` {`title`, `summary`, | |
| 498 | `text`, `annotations`}, `actions`; `PATCH {repo}/check-runs/{id}` to | |
| 499 | complete one). A required check is met by a status or a check run of its | |
| 500 | name. g1t's agents read checks and never report them. Guide: | |
| 501 | https://docs.g1t.sh/guides/checks/ | |
| Secrets and variables: one list, rows per environment, for workflows and deployments | 502 | Secrets and variables are one list per repository (site: |
| 503 | `g1t.sh/<owner>/<repo>/settings/secrets`) and per workspace: each row is a | |
| 504 | key, Secret or Config, the environments it applies to (all, or e.g. | |
| 505 | production/preview, or a job's `environment:`), and whether workflows, | |
| 506 | deployments or both read it. API: `{repo}/actions/secrets` and | |
| 507 | `{repo}/actions/variables` (GitHub's routes) with extra `environments`, | |
| Merge branch 'worktree-agent-a3abfcce648e87dca' | 508 | `available_to`, `repositories`, `note`, `id`. Every job gets |
| 509 | `secrets.G1T_TOKEN` (`GITHUB_TOKEN` is its alias): a token for that job | |
| 510 | only, reaching its repository only, with the scopes its `permissions:` | |
| 511 | give (default `contents: read`, `packages: read`), revoked when the job | |
| 512 | ends; what it changes starts no workflows except `workflow_dispatch` and | |
| 513 | `repository_dispatch` (`POST {repo}/dispatches`). It cannot change | |
| 514 | secrets. A job naming an environment with protection rules (required | |
| 515 | reviewers, wait timer, branch limits; `PUT {repo}/environments/{name}`) | |
| 516 | waits until they pass, and reviewers approve it at | |
| 517 | `POST {repo}/actions/runs/{id}/pending_deployments`. A pull request's | |
| 518 | runs from outside may wait as `action_required` until someone with Write | |
| 519 | approves them (`POST {repo}/actions/runs/{id}/approve`); agents cannot | |
| 520 | approve runs or deployments. A pull request's runs and preview are trusted | |
| Invite-only launch: sign in with GitHub, repository access and lifecycle, many emails, a new look | 521 | only when its author has `write` or higher on the repository (a member or |
| g1t is the stored author of what it opens; the person who asked is requested_by and keeps the author's rights | 522 | an outside collaborator), or is g1t working on its own; for one g1t made, |
| 523 | the role of whoever asked for it (`requested_by`) counts. Anyone else's run | |
| 524 | with config only. Guide: | |
| Secrets and variables: one list, rows per environment, for workflows and deployments | 525 | https://docs.g1t.sh/guides/secrets-and-variables/ |
| Docs: automations | 526 | |
| Fast pages, required checks on the branch, self-hosted runners, honest incidents | 527 | Self-hosted runners: a job with `runs-on: self-hosted` (or |
| 528 | `[self-hosted, linux, gpu]`, or `{group: name}`) waits until a runner of the | |
| 529 | workspace's with every label takes it. Owners add one under | |
| 530 | `g1t.sh/<workspace>/-/runners`: a one-hour registration token, then | |
| 531 | `g1t-runner register --url https://g1t.sh --token g1trt_…` and | |
| 532 | `g1t-runner run`. Runners only connect out. API: | |
| 533 | `/workspaces/{workspace}/actions/runners`, `.../runner-groups`, | |
| 534 | `.../runner-settings` (and the same under `{repo}/actions/` for a | |
| 535 | repository's own); MCP: the `workflow` tool's `list_runners`, | |
| 536 | `create_runner_token`, `remove_runner`, runner group and settings actions | |
| 537 | (`runners:read`/`runners:admin`). Pull requests from forks never run on them | |
| 538 | unless allowed. Guide: https://docs.g1t.sh/guides/self-hosted-runners/ | |
| 539 | ||
| Projects: what a workspace builds and runs, first on every page | 540 | ## Projects |
| 541 | ||
| 542 | A project is what a workspace builds and runs; every repository is a | |
| 543 | project of its own name (`g1t.sh/<owner>/<project>` opens its overview; its | |
| 544 | code is under `/code`; every repository address still works). Deployments, | |
| 545 | secrets and variables belong to the project; branches, pull requests, | |
| 546 | review and merge rules to its repository (Settings → Repository). Guide: | |
| 547 | https://docs.g1t.sh/guides/projects/ | |
| 548 | ||
| Agents get guardrails, run credentials, an audit log, a context hub, repository instructions and mentions; security upkeep; snake_case API | 549 | ## Security |
| 550 | ||
| 551 | A push that adds a known key or token format (AWS, GitHub, GitLab, Stripe | |
| Git storage hardened, pages in tens of milliseconds, honest security alerts, and costs reconciled daily | 552 | live, Slack tokens and webhooks, Google, Anthropic, OpenAI, npm, g1t, |
| 553 | SendGrid, PEM private keys, service-role JWTs) is refused with every secret | |
| 554 | listed by `file:line` in git's output and a link to allow it; this includes | |
| 555 | an agent's push to its pull request. A very large push is scanned after it | |
| 556 | lands, not before: it goes through, its new commits (any branch) are | |
| 557 | scanned in the background, and a secret found is an open alert, emailed to | |
| 558 | the workspace's owners when it looks real. History is scanned once in the | |
| 559 | background. Never commit a secret: read it from the environment. Values are | |
| 560 | judged by the value alone, never the file's path: a documented example key, | |
| 561 | a value containing example/sample/dummy/fake/placeholder/changeme/notreal/ | |
| 562 | redacted/xxxxx, one that counts up (six or more, like 123456), one | |
| 563 | character five or more times, a repeated short piece, or too little | |
| 564 | randomness is a "likely test value": listed apart, never blocks a push, | |
| 565 | never counted critical. Any other fixture carries `g1t:allow-secret` in a | |
| 566 | comment on its line. Alerts are `open`, `dismissed` or `fixed`. Dismissing a | |
| 567 | secret alert needs `admin`, with a reason (`false_positive`, | |
| 568 | `used_in_tests`, `wont_fix`: dismissed, and pushes carrying it go through; | |
| 569 | `revoked`: fixed) and an optional comment (500 characters); a dependency | |
| 570 | alert needs `write` (`fix_started`, `no_bandwidth`, `tolerable_risk`, | |
| 571 | `inaccurate`, `not_used`). Reopen undoes it. API: | |
| 572 | `GET {repo}/security/alerts` (`state`, `kind`), | |
| 573 | `POST {repo}/security/alerts/{id}/dismiss` (`reason`, `comment`), | |
| 574 | `POST {repo}/security/alerts/{id}/reopen`; MCP `repository` actions | |
| 575 | `security_alerts`, `dismiss_alert`, `reopen_alert` (`repo:read` to list, | |
| 576 | `repo:admin` to dismiss or reopen). Lockfiles (npm, pnpm, yarn, Cargo, Go, | |
| 577 | Python) on the default branch are checked against OSV on every push to it | |
| 578 | and daily. Security updates (on by default; `maintain` turns them off): for | |
| 579 | each vulnerable dependency with a fix, g1t itself opens a pull request | |
| 580 | authored by `g1t` from `g1t/security/<package>-<version>`, raising the | |
| 581 | version in each lockfile with the ecosystem's own tool; it merges through | |
| 582 | the branch's required checks and merge queue. A newer update for the same | |
| 583 | package, or the package no longer being vulnerable, closes it as | |
| 584 | superseded. Only when the bump fails, or required checks fail because code | |
| g1t is one name: its agent's work, commits and comments show as @g1t, and nobody can claim g1t or g1t-agent | 585 | must change, does g1t open an issue and assign it to g1t (the session |
| Git storage hardened, pages in tens of milliseconds, honest security alerts, and costs reconciled daily | 586 | shows "started by g1t"). With no fix published, the alert links the |
| 587 | advisory and is checked again daily. `.g1t/dependencies.yml` (version | |
| 588 | updates) is read and validated, but no version update pull requests are | |
| 589 | opened yet. `g1t` is not an account and cannot be signed in to. | |
| Agents get guardrails, run credentials, an audit log, a context hub, repository instructions and mentions; security upkeep; snake_case API | 590 | Guide: https://docs.g1t.sh/guides/security/ |
| 591 | ||
| Teams and CODEOWNERS, labels and milestones, dependency updates, the security suite, and a clearer top bar | 592 | The security suite (scopes `security:read`, `security:write`; MCP tool |
| 593 | `security`). A push refused for a secret links to its alert, where anyone | |
| 594 | with `write` bypasses it with a reason (`false_positive`, `used_in_tests`, | |
| 595 | `will_fix_later`), or asks owners when the workspace delegates bypasses; | |
| 596 | an agent never bypasses: take the secret out and push again. Code | |
| 597 | scanning: upload SARIF 2.1.0 to `POST {repo}/code-scanning/sarifs` | |
| 598 | (`commit_sha`, `ref`, `sarif` gzipped then base64); default-branch results | |
| 599 | become alerts, pull request results (`refs/pull/<n>/head`) become line | |
| 600 | comments and the `Code scanning` status. `Dependency review` is a status on | |
| 601 | every pull request that changes a lockfile. Fix what either reports in your | |
| 602 | own pull request; both can be required checks. Also: custom patterns, | |
| 603 | validity checks, `GET {repo}/dependency-graph/sbom` (SPDX 2.3, in `sbom`), | |
| 604 | `GET {repo}/dependency-graph/compare/{base...head}`, and a workspace | |
| 605 | overview. On private repositories these need the Security and quality | |
| 606 | activation (`402` without it); public repositories have them free. | |
| 607 | Guides: https://docs.g1t.sh/guides/security/secret-protection/, | |
| 608 | https://docs.g1t.sh/guides/security/code-scanning/, | |
| 609 | https://docs.g1t.sh/guides/security/supply-chain/ | |
| 610 | ||
| Deployments: a preview for every pull request, production on g1t.page | 611 | ## Deployments |
| 612 | ||
| Invite-only launch: sign in with GitHub, repository access and lifecycle, many emails, a new look | 613 | Part of the g1t plan ($20 a month per workspace, started by an owner |
| status.g1t.sh with incident management, invites that land you in the workspace, settings as pages, usage without quotas | 614 | under the workspace's Billing). No quotas: unlimited projects and |
| 615 | previews (never charged); builds by the second, requests ($0.36/M), CPU | |
| 616 | ($0.024/M ms) and custom domains ($0.12 a month each) metered from the | |
| 617 | first at cost + 20%, from the plan's $10 first. The trial | |
| Invite-only launch: sign in with GitHub, repository access and lifecycle, many emails, a new look | 618 | never covers deployments. Then someone with admin on the repository |
| Projects: what a workspace builds and runs, first on every page | 619 | turns deployments on for a project (Settings → Deployments, or Deploy on |
| 620 | its overview). Production deploys from the default branch to | |
| 621 | `https://<project>-<owner>.g1t.page` on each push; every branch with an | |
| 622 | open pull request gets a preview at | |
| 623 | `https://<project>-git-<branch>-<owner>.g1t.page` (a fork's pull request is | |
| Git storage hardened, pages in tens of milliseconds, honest security alerts, and costs reconciled daily | 624 | `pr-<n>`), shown on it as the check `g1t / deploy` (`g1t / deploy (<project>)` |
| 625 | for a workspace's other projects). Builds and running apps | |
| Projects: what a workspace builds and runs, first on every page | 626 | read the project's secrets and variables available to Deployments, each |
| 627 | key's Production or Preview row. Workers projects (`wrangler.jsonc`) and | |
| 628 | static sites build without configuration. Previews come down when the pull | |
| Merge branch 'main' into worktree-agent-a69aeabc4b0deeb97 | 629 | request closes and after idle days. |
| Projects: what a workspace builds and runs, first on every page | 630 | Guide: https://docs.g1t.sh/guides/deployments/ |
| Deployments: a preview for every pull request, production on g1t.page | 631 | |
| Merge branch 'main' into worktree-agent-a69aeabc4b0deeb97 | 632 | A repository's deployments, wherever they run, are one list: `source` is |
| 633 | `api` (reported from any CI), `actions` (a g1t Actions job with | |
| 634 | `environment:`) or `g1t_page` (g1t.page builds, ids `dpl_…`, environments | |
| 635 | `production` and `preview`). Read with `GET {repo}/deployments` (filters | |
| 636 | `environment`, `ref`, `sha`, `task`, `state`, `source`, `creator`, `page`, | |
| 637 | `per_page`), `GET {repo}/deployments/{id}`, `GET {repo}/deployments/{id}/statuses`, | |
| 638 | `GET {repo}/environments` and `GET {repo}/environments/{environment}` | |
| 639 | (`deployments:read`, Read role). Report from a CI with | |
| 640 | `POST {repo}/deployments` (`ref`, optional `environment` (default | |
| 641 | `production`), `sha`, `task`, `description`, `payload`, `state`, | |
| 642 | `environment_url`, `log_url`), then | |
| 643 | `POST {repo}/deployments/{id}/statuses` with `state` (`queued`, | |
| 644 | `in_progress`, `success`, `failure`, `error`, `inactive`) and | |
| 645 | `environment_url` (`deployments:write`, Write role). A success makes the | |
| 646 | environment's older successful deployments `inactive` unless | |
| 647 | `auto_inactive: false`. Each status sets the commit check | |
| 648 | `deploy / <environment>`, which a ruleset's `required_deployments` rule can | |
| 649 | require. A g1t Actions job with `environment:` (or `{name, url}`) reports | |
| 650 | by itself, one deployment per run and environment; `deployment: false` | |
| 651 | opts out. MCP: the `workflow` tool's `list_deployments`, `get_deployment`, | |
| 652 | `create_deployment`, `deployment_statuses`, `create_deployment_status`, | |
| 653 | `list_environments`, `get_environment`. Webhooks: `deployment.created`, | |
| 654 | `deployment_status.created`. Guide: https://docs.g1t.sh/guides/deployments-api/ | |
| 655 | ||
| Search across all of g1t, Explore, and a command palette | 656 | ## Search |
| 657 | ||
| Thirteen MCP tools and classic token scopes; agents rate their confidence and can be put on an issue in one step | 658 | `GET https://api.g1t.sh/search?q=<query>&type=<type>` (MCP: `search`, action `code`, the default) |
| Search across all of g1t, Explore, and a command palette | 659 | searches all of g1t: repositories (name, description, topics, README), code |
| 660 | on default branches, issues, pull requests, people and workspaces. No token | |
| Invite-only launch: sign in with GitHub, repository access and lifecycle, many emails, a new look | 661 | needed for public results; with one, private results the token's person |
| 662 | can read are included (their workspaces' repositories, and those they were | |
| 663 | given a role on), checked against current membership and roles. `type` is | |
| Search across all of g1t, Explore, and a command palette | 664 | `repositories`, `code`, `issues`, `pulls` or `people` (worked out from the |
| 665 | qualifiers when left out); `page` and `per_page` (at most 50) page through. | |
| 666 | The query takes words, `"exact phrases"`, `-word` to leave out, and | |
| 667 | `repo:owner/name`, `org:<workspace>`, `language:<lang>`, `path:<prefix or | |
| 668 | *.glob>`, `is:issue`, `is:pr`, `is:open`, `is:closed`, `is:merged`, | |
| 669 | `author:<username>`, `label:<label>`. Code search matches any run of three | |
| 670 | characters or more; vendored directories, lockfiles, binaries and files | |
| 671 | over 512 KB are not indexed. Results give `counts` per type and each hit's | |
| 672 | `snippet` or code `lines` as parts with `highlight`. On the site: | |
| 673 | `https://g1t.sh/search?q=`, ⌘K, and `https://g1t.sh/explore` for public | |
| 674 | projects by activity, language (`?language=`) and topic (`?topic=`). | |
| Thirteen MCP tools and classic token scopes; agents rate their confidence and can be put on an issue in one step | 675 | The `search` tool's `context` action stays the search of one workspace's |
| 676 | context hub. Guide: | |
| Search across all of g1t, Explore, and a command palette | 677 | https://docs.g1t.sh/guides/search/ |
| 678 | ||
| Account dropdown, llms.txt onboarding, hosted agent runner (not yet deployed) | 679 | ## Facts |
| 680 | ||
| Git storage hardened, pages in tens of milliseconds, honest security alerts, and costs reconciled daily | 681 | - API base: `https://api.g1t.sh`. `GET /` lists the main URLs as templates; |
| 682 | every operation is in the OpenAPI document. | |
| API and MCP server in Rust; a public index at the API root | 683 | Auth: `Authorization: Bearer g1t_…`. Public data needs no token. Errors are |
| Account dropdown, llms.txt onboarding, hosted agent runner (not yet deployed) | 684 | `{"error": {"code": "…", "message": "…"}}` with codes `unauthenticated` |
| Git storage hardened, pages in tens of milliseconds, honest security alerts, and costs reconciled daily | 685 | (401), `payment_required` (402, when the plan or a limit refuses compute), |
| 686 | `forbidden` (403, with `needed_scope` when the token lacks a | |
| Thirteen MCP tools and classic token scopes; agents rate their confidence and can be put on an issue in one step | 687 | scope), `not_found` (404), `conflict` (409), `invalid` (422). Each |
| 688 | operation's scope is `x-scope` in the OpenAPI document. The full description is at https://api.g1t.sh/openapi.json. | |
| Workspaces own repositories | 689 | - Git remote: `https://g1t.sh/{workspace}/{repo}.git`. In API paths, |
| Issues and pull requests replace intents and attempts | 690 | `{owner}` is the workspace. Pull request forks: |
| 691 | `https://g1t.sh/pulls/{pull_request_id}.git`. SSH is not available. | |
| What g1t can't do yet, and an open letter to Cloudflare | 692 | - Limits: 1 GB per repository, 32 MB per file, 100 MB per push. Every |
| 693 | current limit, why it exists and the workaround: | |
| 694 | https://docs.g1t.sh/about/limitations/ | |
| Device sign-in replaces registering and minting tokens over the API | 695 | - Forgotten password: https://g1t.sh/forgot (the person does this, in a |
| 696 | browser). | |
| Issues and pull requests replace intents and attempts | 697 | - Times are RFC 3339 in UTC. |
| OAuth 2.1 sign-in for MCP clients and other applications | 698 | - OAuth 2.1 for applications: metadata at |
| 699 | `https://api.g1t.sh/.well-known/oauth-authorization-server`; authorization | |
| 700 | code with PKCE (S256), public clients, dynamic registration. | |
| Fast pages, required checks on the branch, self-hosted runners, honest incidents | 701 | - A pull request whose required checks have not passed (failed, still |
| 702 | running, or not reported yet) is refused a merge with `409`, saying | |
| 703 | which; where the repository allows bypassing them | |
| 704 | (`allow_ignoring_checks`), someone who can merge can send | |
| 705 | `{"ignore_checks": true}`. Checks that are not required never hold a | |
| 706 | merge. With the merge queue on, the workflows behind required checks | |
| 707 | need `merge_group` in their `on:`. | |
| Git storage hardened, pages in tens of milliseconds, honest security alerts, and costs reconciled daily | 708 | - Landing fast-forwards the default branch: g1t makes no merge commit on |
| 709 | main. Bringing a pull request up to date makes a merge commit on its own | |
| 710 | branch. | |
| Invite-only launch: sign in with GitHub, repository access and lifecycle, many emails, a new look | 711 | - Pricing (https://g1t.sh/pricing): the forge is free. One plan, g1t, at |
| 712 | $20 a month per workspace with unlimited members, includes $10 of usage | |
| 713 | at cost + 20% (unused does not roll over). Compute needs the plan or a | |
| 714 | card check (never charged); the $5 trial needs a credit or debit card, | |
| status.g1t.sh with incident management, invites that land you in the workspace, settings as pages, usage without quotas | 715 | not a prepaid one. No quotas on the plan: everything is metered from the |
| 716 | first unit at cost + 20%, and only the spend limit stops work. Every | |
| 717 | workspace has 1 GB of private storage and 50,000 git operations a month | |
| 718 | free; past them, the plan pays ($0.60/GB-month, $0.18/1,000) and a free | |
| 719 | workspace is held (pushes to private repositories stop; git slowed to 60 | |
| 720 | an hour). Limits: $100 in a | |
| Invite-only launch: sign in with GitHub, repository access and lifecycle, many emails, a new look | 721 | paid workspace's first month, rising as payments clear; owners set a |
| 722 | spend limit, prepay, or ask with Raise my limit. Caps: $2 a run and $10 | |
| 723 | an issue by default. A spend spike pauses new compute until an owner | |
| 724 | chooses Keep going or Stop (`/<workspace>/-/billing`). Audit log: 90 days | |
| 725 | on every plan. | |
| Account dropdown, llms.txt onboarding, hosted agent runner (not yet deployed) | 726 | |
| 727 | ## More | |
| 728 | ||
| Device sign-in replaces registering and minting tokens over the API | 729 | - [Quickstart](https://docs.g1t.sh/quickstart/) |
| Docs worth reading, and kept that way | 730 | - [How g1t works](https://docs.g1t.sh/concepts/overview/) |
| Search across all of g1t, Explore, and a command palette | 731 | - [Search and Explore](https://docs.g1t.sh/guides/search/) |
| g1t is one name: its agent's work, commits and comments show as @g1t, and nobody can claim g1t or g1t-agent | 732 | - [g1t's agent](https://docs.g1t.sh/guides/working-with-g1t/) |
| Docs worth reading, and kept that way | 733 | - [Outcomes and plans](https://docs.g1t.sh/guides/outcomes/) |
| 734 | - [Talking to agents](https://docs.g1t.sh/guides/talking-to-agents/) | |
| 735 | - [Bring your own agent](https://docs.g1t.sh/guides/bring-your-own-agent/) | |
| 736 | - [The merge queue](https://docs.g1t.sh/guides/merge-queue/) | |
| 737 | - [Sessions and why-blame](https://docs.g1t.sh/guides/why-blame/) | |
| Device sign-in replaces registering and minting tokens over the API | 738 | - [Forks and branches](https://docs.g1t.sh/concepts/forks/) |
| Docs worth reading, and kept that way | 739 | - [Accounts and sign-in](https://docs.g1t.sh/guides/authentication/) |
| 740 | - [Workspaces and tokens](https://docs.g1t.sh/guides/workspaces/) | |
| Invite-only launch: sign in with GitHub, repository access and lifecycle, many emails, a new look | 741 | - [Access and roles](https://docs.g1t.sh/guides/access-and-roles/) |
| Teams and CODEOWNERS, labels and milestones, dependency updates, the security suite, and a clearer top bar | 742 | - [Teams](https://docs.g1t.sh/guides/teams/) |
| 743 | - [CODEOWNERS](https://docs.g1t.sh/guides/codeowners/) | |
| Invite-only launch: sign in with GitHub, repository access and lifecycle, many emails, a new look | 744 | - [Managing a repository](https://docs.g1t.sh/guides/managing-repositories/) |
| Docs: integrations, and your own model provider | 745 | - [Integrations](https://docs.g1t.sh/guides/integrations/) |
| Model providers: gateway tokens for endpoints, tidier rows, and the docs | 746 | - [Model providers](https://docs.g1t.sh/guides/models/) |
| Merge the AI Gateway: Anthropic's Messages API on a workspace's tokens | 747 | - [AI Gateway](https://docs.g1t.sh/guides/ai-gateway/) |
| Webhooks: every event, to your own addresses, signed and retried | 748 | - [Webhooks](https://docs.g1t.sh/guides/webhooks/) |
| GitHub Actions on g1t, part three: .g1t/workflows, the pages, the docs | 749 | - [GitHub Actions](https://docs.g1t.sh/guides/actions/) |
| Fast pages, required checks on the branch, self-hosted runners, honest incidents | 750 | - [Self-hosted runners](https://docs.g1t.sh/guides/self-hosted-runners/) |
| Docs worth reading, and kept that way | 751 | - [Usage and billing](https://docs.g1t.sh/guides/usage-and-billing/) |
| Docs as their own app; shared theme package | 752 | - [Git](https://docs.g1t.sh/guides/git/) |
| Docs worth reading, and kept that way | 753 | - [MCP tools](https://docs.g1t.sh/reference/mcp/) |
| Merge branch 'worktree-agent-ab2e39e11a6493412' | 754 | - [API reference](https://docs.g1t.sh/reference/api/) |
| What g1t can't do yet, and an open letter to Cloudflare | 755 | - [What g1t can't do yet](https://docs.g1t.sh/about/limitations/) |
| 756 | - [An open letter to Cloudflare](https://docs.g1t.sh/about/open-letter-to-cloudflare/) | |
| Invite-only launch: sign in with GitHub, repository access and lifecycle, many emails, a new look | 757 | - [Source](https://g1t.sh/flagon-io/g1t), MIT licensed |
| 758 | ||
| 759 | ## Help, status and policies | |
| 760 | ||
| status.g1t.sh with incident management, invites that land you in the workspace, settings as pages, usage without quotas | 761 | - [Status](https://status.g1t.sh/): whether each part of g1t is working now, 90 days of uptime, and incidents; the same as JSON at https://status.g1t.sh/status.json |
| Git storage hardened, pages in tens of milliseconds, honest security alerts, and costs reconciled daily | 762 | - [Support](https://g1t.sh/support): where to get help (hey@flagon.io) |
| Invite-only launch: sign in with GitHub, repository access and lifecycle, many emails, a new look | 763 | - [Security](https://g1t.sh/security): how g1t protects code and accounts, and responsible disclosure (hey@flagon.io, https://g1t.sh/.well-known/security.txt) |
| 764 | - [Policies](https://g1t.sh/policies): [Terms of Service](https://g1t.sh/policies/terms), [Privacy Policy](https://g1t.sh/policies/privacy), [Acceptable Use](https://g1t.sh/policies/acceptable-use), [Refunds and Cancellation](https://g1t.sh/policies/refunds), [Subprocessors](https://g1t.sh/policies/subprocessors) | |
| 765 | - g1t is made by Flagon, Inc. (https://www.flagon.io) |
This file's history is long; its oldest lines are credited to the oldest commit read.