agent-browser
Browser automation CLI for AI agents. Use when the user needs to interact with websites, including navigating pages, filling…
Use to merge in main and resolve open merge conflicts.
$ npx -y skills add humanlayer/humanlayer --skill merge-resolve-conflicts --agent claude-codeHow it fires
How this skill gets triggered: by you, by Claude, or both.
/merge-resolve-conflictsContext preview
The summary Claude sees to decide when to auto-load this skill.
Use to merge in main and resolve open merge conflicts.
name: merge-resolve-conflicts description: Use to merge in main and resolve open merge conflicts.
You are either on a branch in a repository, or you are on a git worktree for a specific branch. You are responsible for ensuring the default branch (usually `main` or `master`, but not always) is merged into the current branch.
Begin by running `git fetch`.
Then:
First, you must determine if you are on a worktree for a repository OR if you are on a regular branch on a repository OR if you are on the default branch for the repository
Second, you must determine if the branch you are on has changes on `upstream` / `origin` (as configured) that you do not have on your local branch
Third, you should determine if there are existing merge conflicts on your branch already by running `git status`.
Fourth, you should determine if the user has both an **origin** _and_ an **upstream** configured.
**Important**: if the user is working across multiple repositories, you should ask them which repository(ies) they would like help merging in.
Depending on the results of your orientation you should take one of the following actions:
1. Read the `ticket.md` and structure outline artifacts. These should contain sufficient context about the user's task and what they are working on. You may _later_ read the plan file if you determine that more information is needed. 2. Use sub-agents to analyze the git state on the current branch and on the branch the conflicts come from to understand what was done and why they diverge. 3. Use sub-agents to read the _entire_ file for each conflict, as well as key relevant code and modules to understand the scope and nature of each conflict 4. For each conflict (not necessarily each _individual_ confict, e.g. group two related conflicts in the same file together e.g. for a function change and an import change for it), provide an analysis to the user of: a. What is on HEAD b. What is on the branch that the conflict is coming from (probably `main` or the other default branch) c. Based on the scope of the current task, recent commits to it and to the branch where the source of the conflict is coming from, how you recommend resolving it (keep HEAD, keep other branch, merge them), and why Keep your analysis concise, offer to provide more detail to the user if they ask. 5. Ask the user how they would like to proceed for each one. 6. Work back and forth with the user to resolve all conflicts. Only make changes that the user explicitly asks for - e.g. do not resolve a conflict until the user confirms the approach or instructs you to make the change. 7. Once all merge conflicts are resolved, ask the user if they would like you to commit and push. If they confirm, you should write a conventional commit-style message indicating the branch merged in (from what commit hash), and a full body that indicates what each of the conflicts resolved was and how it was resolved and why.
**IMPORTANT**: Update your todo list to include items 1-7. You _MUST_ stop and ask the user how they would like to proceed after presenting the analysis from step 4. You MUST work back and forth with the user until all conflicts have been resolved before proceeding to stop 7. The user may only give you instructions / guidance on _some_ of the changes at a time - you should use the todo tool to track which ones have been completed and which ones still need resolution in your todo list.
HumanLayer runs coding agents (Claude Code, Codex, OpenCode, Pi) on your machines and lets your team steer them from one place: tasks, sessions, plans, and reviews, synced between a desktop app, a web app, and a daemon on each host. This repo holds all of it.
Browser automation CLI for AI agents. Use when the user needs to interact with websites, including navigating pages, filling…
Rewrites ambiguous English into ASD-STE100 style — one meaning per word, active voice, simple tense, short sentences. Use when…
Use for 1) inspecting PR state, 2) reading review threads, or 3) debugging CI failures. Otherwise use the gh CLI.
Bump the base version of one or more packages (riptide-native, cli, linear-cli, rpi-plugin) on a PR - updates every synced file,…
Diff call stacks across git commits for agentic code review (22 languages). Run `calldiff --help` for usage details.
Use this to skill to write shell scripts or run shell commands to automate tasks in the browser or otherwise use Chrome DevTools…