Team Routing
Productivityteam-routingTeam Routing is an Agent Skill for productivity published in the WellApp-ai/Well repository on GitHub: a SKILL.md instruction file an AI agent can load to do this task.
Detect domain from context and find appropriate team member
How to use this skill
Bring this guide into your coding agent with a prompt tailored to the tool you use.
- Open your project in Codex.
- Copy and paste the prompt below.
- Review files and risks before approving any action.
Inspect the Agent Skill at https://github.com/WellApp-ai/Well/blob/HEAD/cursor-rules/skills/team-routing/SKILL.md and all supporting files. Report provenance, permissions, scripts, dependencies, and security risks. Propose an exact installation into .agents/skills/team-routing, but wait for my approval before writing files or running scripts.
Copying this prompt does not install or run the skill. Review third-party files before use.
Team Routing Skill
Automatically detect domain from branch names or file paths and find the appropriate reviewer.
When to Use
- Push PR mode: auto-assign reviewer
- Init mode: show domain context
- Any mode needing domain owner lookup
Phases
Phase 1: Extract Keywords
Extract domain keywords from context:
From branch name:
feature/tables-filter → keywords: ["tables"]
fix/auth-token-expired → keywords: ["auth"]
chore/settings-refactor → keywords: ["settings"]
From changed files:
git diff --name-only origin/develop | head -20
Match file paths against known patterns:
apps/web/features/tables/→ "tables"apps/api/routes/auth/→ "auth"apps/web/features/settings/→ "settings"
Phase 2: Query Notion Domains Database
API-query-database: [Domains-DB-ID]
Filter: Keywords contains [extracted-keyword]
OR Codebase_Paths contains [file-pattern]
Expected result:
- Domain name
- Topology type (stream-aligned, platform, enabling, complicated-subsystem)
- Primary keywords
Phase 3: Query Team Capabilities
API-query-database: Team Capabilities
Filter: Primary_Domain matches [domain-name]
AND Availability = "Available"
Sort: Open_PR_Count ASC
Expected result:
- Person name
- GitHub handle
- Max Open PRs
- Current Open PR count
Phase 4: Check Availability
For each candidate:
- Query current open PRs:
gh pr list --state open --author [github_handle] --json number | jq length - Compare against Max_Open_PRs threshold
- Skip if at or over capacity
Phase 5: Fallback Logic
If no available primary owner:
- Backup Owners: Query Skills table for overlapping skills
- Skills Match: Match changed file extensions to skill file patterns
- No Match: Return warning: "No available reviewer found for domain [X]"
Output
Present findings as:
## Domain Detection
**Branch:** [branch-name]
**Detected Domain:** [domain-name]
**Topology:** [stream-aligned | platform | enabling | complicated-subsystem]
### Reviewer Assignment
**Primary Owner:** @[github_handle] ([person_name])
**Availability:** [Available | At capacity (X/Y PRs)]
**Assignment:** [Assigned | Fallback to @backup | No available reviewer]
Integration
This skill is invoked by:
push-pr.mdc- Phase 1.2 (Find Reviewer)init.mdc- Domain context display
Notion Databases Required:
- Domains (keywords, codebase paths)
- Team Capabilities (person, domain, availability, max PRs)
- Skills (file patterns for fallback matching)