mlflow / copilot
Install for your project team
Run this command in your project directory to install the skill for your entire team:
mkdir -p .claude/skills/copilot && curl -L -o skill.zip "https://fastmcp.me/Skills/Download/2990" && unzip -o skill.zip -d .claude/skills/copilot && rm skill.zip
Project Skills
This skill will be saved in .claude/skills/copilot/ and checked into git. All team members will have access to it automatically.
Important: Please verify the skill by reviewing its instructions before using it.
Hand off a task to GitHub Copilot.
0 views
0 installs
Skill Content
---
name: copilot
description: Hand off a task to GitHub Copilot.
allowed-tools:
- Bash(gh agent-task create:*)
- Bash(gh agent-task list:*)
- Bash(gh agent-task view:*)
- Bash(bash .claude/skills/copilot/poll.sh *)
- Bash(gh api:*)
---
## Examples
```bash
# Create a task with an inline description
gh agent-task create "<task description>"
# Create a task from a markdown file
gh agent-task create -F task-desc.md
```
## Post-creation
Print both the session URL and the PR URL (strip `/agent-sessions/...` from the session URL).
Example:
- Session: https://github.com/mlflow/mlflow/pull/20905/agent-sessions/abc123
- PR: https://github.com/mlflow/mlflow/pull/20905
## Polling for completion
Once Copilot starts working, poll in the background until Copilot finishes. The script automatically finds the latest session for the PR:
```bash
bash .claude/skills/copilot/poll.sh "<owner>/<repo>" <pr_number>
```
## Sending feedback
If the PR needs changes, batch all feedback into a single review with `@copilot` in each comment so they're addressed in one session:
```bash
gh api repos/<owner>/<repo>/pulls/<pr_number>/reviews --input - <<'EOF'
{
"event": "COMMENT",
"comments": [
{
"path": "<file_path>",
"line": <line_number>,
"side": "RIGHT",
"body": "@copilot <comment>",
// ... more params
},
// ... more comments
]
}
EOF
```