- Home
- Skills
- Agents & Orchestration
- One Task, One Branch Handoff
One Task, One Branch Handoff
Each task gets its own branch and a short handoff note, checked before you stop, so the next session starts from facts.
$40
Works with the AI Agents you already use
One Task, One Branch Handoff
Example session with this skill installed
Add a function mul(a, b) to app.py that returns a times b, with a test. I am stopping for today and will continue tomorrow, so leave things so I can pick it up.
- Read your context and instructions
- Compiled the one task one
Summary
The agent started a task branch, did the one task, wrote the handoff note, committed it, and ran the check. It did not merge or push. Output below is from running the script on a small test repo.
Start
$ python3 scripts/handoff.py new add-mul
branch task/add-mul
note .handoff/add-mul.md
The finished note, .handoff/add-mul.md
# Handoff: add-mul
## Goal
Add mul(a, b) to app.py, returning a times b.
## Done
- mul added to app.py with a test in tests/test_app.py
## Not done
- nothing
## Decisions
- A plain function, because the task asked for one function.
## Next step
Review the diff and merge task/add-mul.
## How to check
- `python3 -m unittest`
Check before stopping
$ python3 scripts/handoff.py check
PASS branch: task/add-mul
PASS note: .handoff/add-mul.md
PASS note committed
PASS tree: clean
Exit code 0. An unfinished template note fails with one line per empty section, for example FAIL note: empty section: Goal, and the check exits 1.
Next session
The agent reads the note first and runs the "How to check" command, so it starts from the real state and not from old chat.
Connects securely to your tools. The creator never sees your data.
About this skill
Chat context slips. A branch and a note do not.
When a task starts, the agent creates a branch named task/SLUG and a handoff note from a template. It works on that one goal and writes any second job into the note instead of starting it.
When the task or session ends, the agent fills six sections: Goal, Done, Not done, Decisions, Next step, How to check. A script then checks the work: the branch name, a complete note, the note committed on the branch, a clean tree. It fails if you are on main or a detached HEAD.
Next session, the agent reads the note first and runs the "How to check" commands to learn the real state.
Standard library only. Written for Python 3.10 or later, tested on 3.13 only, git needed. The script has 13 tests. It works from anywhere inside the repo.
What it does not do It does not merge, push or open a pull request. It checks that the note is complete and committed, not that it is true. It does not stop edits outside the task.
Before you use this These skills are new and have not been tested widely. They come from a project that works and that I use. You are responsible for what you run and for reviewing what your agent does with them, so check results before you rely on them. The author is not liable for loss or damage, in line with the MIT licence.
Requires Python 3 and Git. Does not automate merges or PR creation. Note validation checks for completion and commit status, not technical accuracy.
How to use
- 1
Connect your agent to the Agensi MCP
Copy the prompt and paste it into Claude, ChatGPT, Cursor or any agent that supports MCP. It sets up the connection for you. Manual setup
- 2
Sign in to Agensi
Your agent shows a sign-in link. Sign in and click Allow.
- 3
Ask for what you need
Your agent loads this skill and uses it on your task. It always asks before any payment.
Reviews
No reviews yet
Be one of the first to try it. Every listed skill goes through the trust checks below.
Security scanned
Checked by our 8-point scan before listing
Fresh listing
Recently published to Agensi
Money-back guarantee
Doesn't do what its page promises? You get your money back.
Trust & safety
Security scan
Security scan passed today
Requires
Permissions
No network access. It runs git commands, creates a branch, and writes a note under .handoff/ in your repo.