Skip to content
⌘K
Explore
Download
Docs
Sign in
Sign up
flagon-io
/
g1t
public
Star
0
Overview
Code
Issues
Pull requests
Workflows
Deployments
Insights
Files
Commits
Branches
Tags
Releases
Compare
ef55ddfeacce171116081002a666e3b7d467f291
g1t
/
services
syntaqx
Merge branch 'worktree-agent-aef0bb4ceda89e542'
ef55ddf
18h ago
History
..
actions
artifacts
billing
chat
deployments
events
identity
integrations
notify
og
packages
pages
projects
repos
runner
search
security
telemetry
webhooks
work
..
actions
A file's history and a repository's failed runs are one read each, so Ask (and any agent) answers "why does this file exist" and "why did the last Deploy fail" from what happened rather than by paging. list_commits with `path` now gives that file's or directory's own history: the commits on the first-parent chain that changed it (added, changed, removed), newest first, with `last_changed` the newest and `complete` false when a long history was not walked to its start in time. The repos service's new path_log walks the history as last_commits does, reading ahead in batches, a tree only where the path's directory changed, within its read budget and 8 s. list_workflow_runs takes GitHub's `status`, a status (queued, in_progress, waiting, completed) or a conclusion (failure, success, cancelled…), on REST (status, and head_sha on a workflow's runs too) and MCP, and the actions service filters by either. Ask's system prompt finds failed runs with status failure, says a run still going has no result rather than calling it passed, judges health from the newest completed run and what is still going, and reads a file's history for why it exists, quoting its commit messages. The quality bar no longer reads a run's number, in a link to the run or a table of runs, as an issue's. Its last run on production scored 97.1% of checks and a judge mean of 3.81 over 27 questions, against 52.1% and 3.08 before Ask had tools; these address its weakest answers. Checked: path_log's walks over a file, a directory, a missing path, a limit, pages and a read budget; cargo test for the workspace (the actions crates on their own), the wasm builds, the TypeScript typecheck, every workspace's tests, the deploy tool's and the plugin's tests and the docs build; the OpenAPI document, the CLI's operations and its reference are regenerated, and the MCP reference says both.
20h ago
artifacts
g1t coordinates the work and the person's own sessions do it: Ask g1t reads g1t with the same tools any agent gets, answers in one conversation shared with the @g1t direct message in Chat, and hands work to sessions in the person's own harness, whose subagents g1t records, ties to issues and pull requests, and lets see each other.
23h ago
billing
A workspace runs 20 sandboxes at once on the free plan, the trial and in its first month, and 40 after, where it ran 2 and 10: those caps were sized for g1t's former hosted agents, and held a workflow's jobs, the merge queue and deploys behind one another. A staff override still wins.
22h ago
chat
Ask answers "what shipped today" for the asker's own today, and says what changed rather than only how much. "Today" begins at the asker's local midnight: integrations reads the time zone on their profile, or failing that the one their browser sends with the question (the Ask panel and the Chat composer send it; chat keeps it on a question to @g1t, an IANA name only), works out with the runtime's Intl when their day began in UTC (local_day.rs, tested across zones west and east of UTC and on a half hour), and the system prompt says "they are in America/Denver, where it is 2026-10-12: today means since 2026-10-12T06:00:00Z"; without a zone it stays UTC. get_workspace_activity's answer gains `landed`: for the repositories whose default branch was pushed to in the span (busiest first, at most eight), the commits made in the span, one line each (repo, short sha, subject, author, time, link), at most forty, so a workspace that pushes to main rather than merging pull requests still shows what shipped; the prompt has Ask say what those changes were, pass `since` as the start of the asker's today, and narrow the kinds or the span when `complete` is false rather than guessing. A question asked in Chat keeps the direct message's path when its page has none of its own. The quality bar (bench/ask) asks through the direct message with @g1t, as the panel and Chat do now: it opens it, marks a new conversation, posts each question and follows g1t's answer to its end, and grades the steps, sources and proposals it kept. Its first run on production, with tools, had shipped-today pass every check (8/8) and the judge mark it 2.83, for the UTC day and the thin summary this fixes. Checked: cargo test for the workspace (the actions crates' repository-file tests on their own, as on this machine they cannot find their files from the workspace), the wasm build of integrations and the API, the TypeScript typecheck, every workspace's tests, the deploy tool's and the plugin's tests, and the OpenAPI document regenerated.
21h ago
deployments
A preview is live only while what it previews is: a branch's preview (one with no pull request) comes down when its branch is deleted, and the sweep takes down any whose branch is already gone. A pull request's preview has always come down when it closes or merges, but a branch's had nothing to take it down, so the previews that preview stacks built for another project's pull request, before projects stopped depending on each other (a9916137), stayed up and showed as a live Preview on projects with no pull requests at all, as flagon-io/automation-lab's v2 did. Moving a project no longer builds such a preview again (rebuildStack is gone with the stacks). A branch's preview from a project's deploy workflow stays while its branch is there, and one whose branches can't be read stays until they can. Checked: the deployments service's typecheck and its 88 tests (new: the sweep takes down a preview of a deleted branch and marks its deployment inactive while keeping a pull request's and a live branch's, and git.deleted takes a branch's preview down at once).
18h ago
events
g1t coordinates the work and the person's own sessions do it: Ask g1t reads g1t with the same tools any agent gets, answers in one conversation shared with the @g1t direct message in Chat, and hands work to sessions in the person's own harness, whose subagents g1t records, ties to issues and pull requests, and lets see each other.
23h ago
identity
Merge branch 'worktree-agent-a6da31375506fc440'
19h ago
integrations
A file's history and a repository's failed runs are one read each, so Ask (and any agent) answers "why does this file exist" and "why did the last Deploy fail" from what happened rather than by paging. list_commits with `path` now gives that file's or directory's own history: the commits on the first-parent chain that changed it (added, changed, removed), newest first, with `last_changed` the newest and `complete` false when a long history was not walked to its start in time. The repos service's new path_log walks the history as last_commits does, reading ahead in batches, a tree only where the path's directory changed, within its read budget and 8 s. list_workflow_runs takes GitHub's `status`, a status (queued, in_progress, waiting, completed) or a conclusion (failure, success, cancelled…), on REST (status, and head_sha on a workflow's runs too) and MCP, and the actions service filters by either. Ask's system prompt finds failed runs with status failure, says a run still going has no result rather than calling it passed, judges health from the newest completed run and what is still going, and reads a file's history for why it exists, quoting its commit messages. The quality bar no longer reads a run's number, in a link to the run or a table of runs, as an issue's. Its last run on production scored 97.1% of checks and a judge mean of 3.81 over 27 questions, against 52.1% and 3.08 before Ask had tools; these address its weakest answers. Checked: path_log's walks over a file, a directory, a missing path, a limit, pages and a read budget; cargo test for the workspace (the actions crates on their own), the wasm builds, the TypeScript typecheck, every workspace's tests, the deploy tool's and the plugin's tests and the docs build; the OpenAPI document, the CLI's operations and its reference are regenerated, and the MCP reference says both.
20h ago
notify
Chat, artifacts and notifications no longer serve g1t's own agents. Direct messages and conversation invites take people only, and a space or folio is shared with people and teams. Chat stops waking agents, drops the RPCs agents posted, read, searched and handed off through, and stops answering buttons on agent cards; old cards keep their links. An agent's old messages still show, under its saved name or as "A deleted agent". Artifacts drops the folio and docs RPCs agents read and edited pages with, and an old agent member of a space can still be removed. Notifications drop "an agent is waiting on you". A deployment no longer trusts a pull request with secrets just because g1t's agent opened it, and the sign-up card no longer promises a team of agents.
yesterday
og
Chat, artifacts and notifications no longer serve g1t's own agents. Direct messages and conversation invites take people only, and a space or folio is shared with people and teams. Chat stops waking agents, drops the RPCs agents posted, read, searched and handed off through, and stops answering buttons on agent cards; old cards keep their links. An agent's old messages still show, under its saved name or as "A deleted agent". Artifacts drops the folio and docs RPCs agents read and edited pages with, and an old agent member of a space can still be removed. Notifications drop "an agent is waiting on you". A deployment no longer trusts a pull request with secrets just because g1t's agent opened it, and the sign-up card no longer promises a team of agents.
yesterday
packages
The work, events, webhooks, actions, repos and packages services no longer run g1t's own agents. Work drops the 35 calls that queued issues for g1t, delegated them, messaged and reviewed through the agent, planned, captured memories automatically, rated confidence and capped agent spend, and its pull request lifecycle no longer revises or reviews: a pull request only advances to a merge by the repository's auto-merge setting. Merging a pull request that is behind now catches it up itself, through the repository's branch update, and lands it when the push arrives; when it can't be caught up cleanly the merge says to merge the base in and try again. Recording a session on a pull request, memory recall and remember, overlaps, the merge queue and required checks stay as they were.
yesterday
pages
Projects have Observability: their requests, logs, traces, metrics and errors, from apps on g1t.page with nothing to set up, and from software anywhere else over OpenTelemetry.
yesterday
projects
Merge branch 'worktree-agent-a5b0d9e2733e9e382'
yesterday
repos
A file's history and a repository's failed runs are one read each, so Ask (and any agent) answers "why does this file exist" and "why did the last Deploy fail" from what happened rather than by paging. list_commits with `path` now gives that file's or directory's own history: the commits on the first-parent chain that changed it (added, changed, removed), newest first, with `last_changed` the newest and `complete` false when a long history was not walked to its start in time. The repos service's new path_log walks the history as last_commits does, reading ahead in batches, a tree only where the path's directory changed, within its read budget and 8 s. list_workflow_runs takes GitHub's `status`, a status (queued, in_progress, waiting, completed) or a conclusion (failure, success, cancelled…), on REST (status, and head_sha on a workflow's runs too) and MCP, and the actions service filters by either. Ask's system prompt finds failed runs with status failure, says a run still going has no result rather than calling it passed, judges health from the newest completed run and what is still going, and reads a file's history for why it exists, quoting its commit messages. The quality bar no longer reads a run's number, in a link to the run or a table of runs, as an issue's. Its last run on production scored 97.1% of checks and a judge mean of 3.81 over 27 questions, against 52.1% and 3.08 before Ask had tools; these address its weakest answers. Checked: path_log's walks over a file, a directory, a missing path, a limit, pages and a read budget; cargo test for the workspace (the actions crates on their own), the wasm builds, the TypeScript typecheck, every workspace's tests, the deploy tool's and the plugin's tests and the docs build; the OpenAPI document, the CLI's operations and its reference are regenerated, and the MCP reference says both.
20h ago
runner
Every sandbox whose traffic is intercepted trusts the egress certificate: once a sandbox has any outbound handler, Cloudflare passes all its traffic through the runner Worker and re-signs HTTPS, so an open sandbox with only the abuse-report host intercepted got SELF_SIGNED_CERT_IN_CHAIN on every HTTPS request after the network guardrails were removed (before, g1t's own sandboxes were restricted, and restricted ones were given the certificate).
22h ago
search
Merge main into rm-agents: mirroring removal, open signups, free-tier limits, the g1t CLI and runner fixes, with the agent removal kept. The CLI's table of operations is regenerated from the API without the agent, customize and AI credit operations; no CLI command used them.
yesterday
security
Merge main into rm-agents: mirroring removal, open signups, free-tier limits, the g1t CLI and runner fixes, with the agent removal kept. The CLI's table of operations is regenerated from the API without the agent, customize and AI credit operations; no CLI command used them.
yesterday
telemetry
Merge main into rm-agents: mirroring removal, open signups, free-tier limits, the g1t CLI and runner fixes, with the agent removal kept. The CLI's table of operations is regenerated from the API without the agent, customize and AI credit operations; no CLI command used them.
yesterday
webhooks
The contracts no longer describe g1t's own agents, and workspaces can keep their AI providers. Every consumer had already stopped using them, so the TypeScript and Rust contracts drop the workspace-agent, ability, look, skill-library, context, runner-agent and mention modules, the chat and docs RPCs agents posted and read through, the folio agent methods, the agent_waiting notification, agent.asked and memory.changed events, AI credit, token and gateway billing types, model routes and sessions, team agents, agent tokens, plan, review, message and confidence types, command rules, cost caps and confidence rules. The agents:run and models:* scopes are no longer offered by any preset, OAuth default or permission form, but a stored token or grant that names them still reads; and the removed operations are gone from the scope table, so the API's check that every operation has exactly one scope is strict again. What stored data needs in order to parse (old run and credential kinds, an agent team lead, retired rules and webhook events, old guardrail and integration settings) still parses, and a run's token is still PrincipalKind::Agent, which now means a run's token and the @g1t rows written before. The customize, marketplace, skills and claude-plugin contracts stay only until the mobile app drops its Customize screens.
yesterday
work
Merge branch 'worktree-agent-aef0bb4ceda89e542'
18h ago
