context-partitioning
Context Partitioning — Manifest-First Dispatch
The Problem
Without explicit context isolation, the main orchestrator context window accumulates:
- Full file reads that were only needed once
- Subagent outputs that are already committed to disk
- Skill files that don't apply to the current task
- Historical reasoning that compaction will destroy anyway
Result: context bloat → compaction → rule violations → wrong output.
The Solution: Manifest-First
Rule: Before dispatching ANY subagent, the orchestrator MUST produce a typed manifest. The manifest is the subagent's entire world — it reads NOTHING outside its manifest.
More from cleanexpo/unite-hub
tdd
Use when implementing any feature or bug fix. Hard gate — no production code without a failing test first. Applies to vitest (apps/web/) and pytest (apps/backend/). Triggers on "implement", "add feature", "fix bug", "new component", "new endpoint", or any code-writing task.
1idea-to-production
Plain-English pipeline from idea to production — routes user requests to the right phase and agent
1oauth-flow
OAuth 2.0 and OIDC integration with PKCE, Supabase Auth providers, and redirect URI validation
1health-check
>-
1csv-processor
>-
1vector-search
>-
1