PR Import Workflow

The framework supports creating tasks from external pull requests, providing a structured path for incorporating contributions that cannot be merged directly. Instead of accepting or rejecting a PR as-is, the workflow extracts the contributor’s intent and approach, validates it against project conventions, and produces a task that goes through the standard implementation cycle — with proper attribution back to the original contributor.

Motivation

Not every pull request can be merged directly. Common scenarios:

• External contributions need additional review, testing, or alignment with project conventions before the changes can be accepted
• Code quality gates require that all changes go through the project’s planning and review workflow, regardless of source
• Approach validation — a PR may have the right idea but an implementation that doesn’t fit the codebase patterns, requires refactoring, or has missing edge cases

Importing PRs as tasks solves this by:

• Preserving the contributor’s ideas and implementation details as structured context for a new task
• Running the changes through the same planning → implementation → review cycle as any other task
• Crediting the original contributor via Co-authored-by trailers in the final commits
• Automatically closing the source PR with implementation notes once the task is complete

How It Works

There are two paths for creating a task from a pull request:

Path 1: AI-Powered Review (/aitask-pr-import)

The /aitask-pr-import skill provides a fully interactive, AI-assisted workflow:

1. Select a PR (by number, browse open PRs, or use previously extracted data)
2. AI analyzes the PR — purpose, approach, quality, concerns, codebase alignment
3. Interactive Q&A to explore specific aspects
4. Discover and fold related pending tasks
5. Create a task with rich context and AI-generated implementation approach

This is the recommended path for most PRs, especially those requiring careful analysis.

Path 2: Direct Import (ait pr-import)

The ait pr-import command provides direct import without AI analysis. It supports both interactive and batch modes:

Interactive mode (default) — browse open PRs with fzf, preview details, select which to import, and choose between creating a basic task or extracting data for later use with /aitask-pr-import:

ait pr-import                                         # Interactive PR selection

Batch mode — for automation and scripting:

ait pr-import --batch --pr 42                         # Import a single PR as a task
ait pr-import --batch --pr 42 --data-only             # Extract data only (for /aitask-pr-import)
ait pr-import --batch --all --skip-duplicates         # Import all open PRs

This creates tasks with the PR title, description, and metadata — but without the AI analysis that /aitask-pr-import provides. Use this for bulk imports or when the PR is straightforward.

Both paths store the same metadata in the task frontmatter: pull_request (URL), contributor (username), and contributor_email (for attribution).

End-to-End Flow

External PR
    │
    ├─── /aitask-pr-import ──── AI analysis ──── Task + Plan
    │                                                │
    └─── ait pr-import ────────────────────────── Task (basic)
                                                     │
                                              /aitask-pick
                                                     │
                                          Planning → Implementation
                                                     │
                                          Review → Commit (with Co-authored-by)
                                                     │
                                          Archive → PR closed with notes

Automated Lifecycle

Once a task is created from a PR, the standard task workflow (invoked by /aitask-pick and related skills) handles the rest automatically:

Contributor attribution — During the commit step (Step 8 of task-workflow), if the task has contributor and contributor_email metadata, the commit message includes a Co-authored-by trailer crediting the original contributor:

feature: Implement auth token refresh (t83)

Co-authored-by: contributor-name <contributor-email@users.noreply.github.com>

Platform-specific email formats:

• GitHub: <id>+<username>@users.noreply.github.com (resolved from GitHub API)
• GitLab: <username>@users.noreply.gitlab.com
• Bitbucket: Resolved from the PR author’s profile

PR close/decline — During archival (Step 9 of task-workflow), if the task has a pull_request field, the workflow offers to close or decline the source PR with a comment containing:

• A reference to the resolved task
• Final implementation notes from the archived plan file
• Associated commits detected from git history
• A thank-you note to the contributor

The close behavior is platform-specific: GitHub and GitLab close the PR/MR, Bitbucket declines it.

Task Metadata Fields

Tasks created from PRs include these additional frontmatter fields:

These fields are preserved through the full task lifecycle and used during commit attribution and PR close/decline.

Platform Examples

GitHub

# Extract PR data for AI review
ait pr-import --batch --pr 42 --data-only --silent

# Review with AI
/aitask-pr-import    # Select "Use existing PR data", pick #42

# After implementation and archival:
# → Commit includes: Co-authored-by: user <id+user@users.noreply.github.com>
# → PR #42 is closed with implementation notes

GitLab

# Import MR directly as a task
ait pr-import --batch --pr 15 --source gitlab

# For cross-repo MRs:
ait pr-import --batch --pr 15 --source gitlab --repo group/project

Bitbucket

# Import from Bitbucket Cloud
ait pr-import --batch --pr 7 --source bitbucket

# After archival: PR is declined (Bitbucket uses "decline" semantics)