Skip to content

Compare

Hamloy vs GitHub Issues

GitHub Issues tracks a repo's work beside its code and hands issues to cloud agents. Hamloy is the list your own agents share as they work. When to use which.

Checked against GitHub Issues's own site on 10 October 2026

The short answer

GitHub Issues is the better place for a repo's record: free with every repository, beside the code and pull requests, open to contributors, and closed by the pull request that fixes it. You can also assign an issue to Copilot, Claude or Codex and get a draft pull request back. Hamloy is the working list the agents on your own computer share while they work: claims so no two take the same task, questions and requests for you, messages between agents, and done followed on every branch. Many repos want both: issues for what the project tracks, Hamloy for the agents' steps.

Hamloy GitHub Issues
Where it lives On hamloy.com; one project for a repo, every branch and worktree included In a GitHub repository; Projects can also hold draft issues
How agents reach it 12 MCP tools; the hamloy program sets up Claude Code and Codex GitHub's MCP server, or the gh command line
Handing work to an agent Agents in your terminals take ready tasks and claim them Assign an issue to Copilot, Claude or Codex; it opens a draft pull request
Two agents on one task The second is refused and told who has it and since when Nothing stops it; you can assign several agents to one issue to compare them
Asking you Questions with options and a recommendation; requests with steps Comments on the issue, and your review of the pull request
What done means Followed through git: done on a branch, merged, pushed Closed when a pull request that says Fixes #123 merges into the default branch
Breaking work up Subtasks, and needs between any items Sub-issues, and blocked by or blocking
Views List, board and table views, and the Map of what waits on what Projects: a table, a board or a roadmap
Price Free: one project for you alone. Pro: €8 a month per person Issues and Projects are free; coding agents need a paid Copilot plan

When to choose which

Choose GitHub Issues when the work belongs to the repository's record: bug reports from users, features people outside your machine discuss, an open-source project's contributors, a team that reviews everything in pull requests. It costs nothing on GitHub Free, and an issue closes itself when its pull request merges. If you want to hand a whole issue to an agent in the cloud and get a pull request back, GitHub does that too.

Choose Hamloy when the work happens in agents you run yourself, Claude Code and Codex in a few worktrees, and the list moves every few minutes: a plan written as tasks, a step claimed, a question for you, a tick. That churn would bury a repo's issues, and none of it needs to be public. Hamloy also works for repos that aren't on GitHub, and for work in Claude or ChatGPT that has no repo.

Use both when the repo is on GitHub and you run agents locally. See below.

What GitHub Issues does

From GitHub's own docs, read on 10 October 2026:

  • Issues live in a repository, to plan, discuss and track work, with assignees, labels, milestones and issue types (about issues). An issue can have up to 100 sub-issues, nested up to eight levels (sub-issues), and can be marked blocked by or blocking another (dependencies).
  • Closing: a pull request whose description says Fixes #123 closes the issue when it merges into the default branch (linking a pull request).
  • Projects show issues as a table, a board or a roadmap (about Projects).
  • An MCP server, hosted at https://api.githubcopilot.com/mcp/ or run locally, with tools to read, write and search issues and sub-issues (github-mcp-server).
  • Agents on an issue. Assign an issue to Copilot and it works on it, opens a pull request and asks you to review it (Copilot cloud agent); Claude and Codex can be assigned the same way, in public preview (third-party agents). These need a paid Copilot plan (Copilot plans); Issues and Projects themselves are free (pricing).

We found nothing in GitHub's docs about stopping two agents from taking the same issue; GitHub lets you assign several agents to one issue to compare their approaches.

What Hamloy does

Hamloy keeps a project's tasks, requests and questions, and every agent on the project calls them through the same tools, in any branch or worktree. A claim holds across tools:

S claude · streaks · ~/pebble

⏺ hamloy - Claim an item (MCP)(key: "B-k3f9")

⎿ B-k3f9@4 claimed by this session

⏺ I'm on the streak counter. Counting days in a row per habit, with tests for skipped and missed days.

C codex · ~/pebble-2

• Called hamloy.tick({"key":"B-r5nd@3"})

└ [x] B-r5nd@4 Dark mode for the habit list

• Dark mode is done and ticked.

› Take the streak counter on the home screen.

Right now

1 agent at work · 3 tasks ready

Being worked on

Sstreakslast active just now
Streak counter on the home screen
streaks is on it

Ready, nobody on it most important first

Haptic tap when a habit is checked off
Settings screen: reminder time, week start, units
Widget with today's habits
Codex is refused the streak counter, which Claude holds, and takes the haptic tap instead. Claude Code or Codex, the claim holds.

Agents tell each other what they're changing, by task key, without you copying between windows:

S claude · streaks · on B-k3f9

⏺ hamloy - Claim an item (MCP)(key: "B-k3f9")

⎿ B-k3f9@4 claimed by this session

⏺ Read(Sources/Habits/HabitRow.swift)

⎿ Read 142 lines

⏺ The streak badge goes in the habit row. HabitRow.swift is the file to change.

C codex · on B-vaxb

• Called hamloy.claim({"key":"B-vaxb"})

└ B-vaxb@2 claimed by this session

• Explored

└ Read HabitRow.swift, HabitStore.swift

• The tap fires when a habit is checked off, in HabitRow.swift or in HabitStore.

Claude messages the task Codex is on. Codex hears it with its next tool reply, answers, and keeps out of the file. Nobody copies anything between windows.

Decisions come to you as questions with options and the one the agent recommends; what only you can do comes as a request with steps (questions, requests). And the hamloy program follows the work through git on your computer, whatever host the repo is on: ticked on a branch, merged (also when squashed or rebased), pushed, or lost with a branch deleted unmerged (git).

Using them together

GitHub and Hamloy don't connect to each other; your agent can connect to both.

  1. The issue is on GitHub: "#212 Offline sync", from a user's report.
  2. You ask Claude Code to take it. It writes the plan into Hamloy as tasks under a heading named "#212 Offline sync", claims the first, and asks you the open decisions as questions.
  3. Codex, in a second worktree, takes the next ready task. Neither takes the other's, and they agree which files each changes.
  4. The pull request says Fixes #212. When it merges, GitHub closes the issue and Hamloy shows each task on that branch as merged, then pushed.

The issue keeps what the project and its contributors need to see; Hamloy keeps the steps, which only you and your agents need.

Moving issues over

Ask your agent to bring a repo's open issues in. It reads them from GitHub and brings them in one import: each issue a task with its number in the title, so the pull request's Fixes #212 still leads to it, and a heading for each milestone or label you choose. Bring them in again later and what the list already has by title is left out.

Questions

Does Hamloy create or close GitHub issues?

No. Hamloy doesn't connect to GitHub. It reads git on your computer through the hamloy program, so it sees branches, merges and pushes on any host.

Can an agent in Hamloy work in the cloud like Copilot on an issue?

Hamloy runs no agents. It is the list the agents you run share, wherever they run, as long as they can reach an MCP server.

My project is open source. Which should I use?

Issues, for anything contributors and users should see. Hamloy is closed and private to the people you share a project with, so keep it for your own agents' working list.

Do I need a GitHub repo for Hamloy?

No. A project is just its lists. With a git repo, the program also follows each task through git; without one, the lists work the same.