Iterative Retrieval Pattern

Solves the "context problem" in multi-agent workflows where subagents don't know what context they need until they start working.

The Problem

Subagents are spawned with limited context. They don't know:

•Which files contain relevant code
•What patterns exist in the codebase
•What terminology the project uses

Standard approaches fail:

•Send everything: Exceeds context limits
•Send nothing: Agent lacks critical information
•Guess what's needed: Often wrong

The Solution: Iterative Retrieval

A 4-phase loop that progressively refines context:

code

┌─────────────────────────────────────────────┐
│                                             │
│   ┌──────────┐      ┌──────────┐            │
│   │ DISPATCH │─────▶│ EVALUATE │            │
│   └──────────┘      └──────────┘            │
│        ▲                  │                 │
│        │                  ▼                 │
│   ┌──────────┐      ┌──────────┐            │
│   │   LOOP   │◀─────│  REFINE  │            │
│   └──────────┘      └──────────┘            │
│                                             │
│        Max 3 cycles, then proceed           │
└─────────────────────────────────────────────┘

Phase 1: DISPATCH

Initial broad query to gather candidate files:

code

Start with high-level intent:
  patterns: src/**/*.cpp, include/**/*.hpp
  keywords: relevant domain terms
  excludes: *_test.cpp, *_bench.cpp

Phase 2: EVALUATE

Assess retrieved content for relevance:

Scoring criteria:

•High (0.8-1.0): Directly implements target functionality
•Medium (0.5-0.7): Contains related patterns or types
•Low (0.2-0.4): Tangentially related
•None (0-0.2): Not relevant, exclude

Phase 3: REFINE

Update search criteria based on evaluation:

•Add new patterns discovered in high-relevance files
•Add terminology found in codebase
•Exclude confirmed irrelevant paths
•Target specific gaps

Phase 4: LOOP

Repeat with refined criteria (max 3 cycles).

Stop when:

•3+ high-relevance files found
•No critical gaps remain
•Max cycles reached

Practical Examples

Example 1: Bug Fix Context

code

Task: "Fix the memory leak in the particle solver"

Cycle 1:
  DISPATCH: Search for "particle", "solver", "allocat" in src/**
  EVALUATE: Found particle_solver.cpp (0.9), memory_pool.hpp (0.8), main.cpp (0.3)
  REFINE: Add "arena", "pool" keywords; exclude main.cpp

Cycle 2:
  DISPATCH: Search refined terms
  EVALUATE: Found arena_allocator.hpp (0.95), smart_ptr_utils.hpp (0.85)
  REFINE: Sufficient context (4 high-relevance files)

Result: particle_solver.cpp, memory_pool.hpp, arena_allocator.hpp, smart_ptr_utils.hpp

Example 2: Feature Implementation

code

Task: "Add MPI communication to the mesh partitioner"

Cycle 1:
  DISPATCH: Search "MPI", "mesh", "partition" in src/**
  EVALUATE: No matches for "MPI" - codebase uses "comm" namespace
  REFINE: Add "comm::", "distribute", "scatter" keywords

Cycle 2:
  DISPATCH: Search refined terms
  EVALUATE: Found comm_layer.hpp (0.9), mesh_partition.cpp (0.7)
  REFINE: Need data serialization patterns

Cycle 3:
  DISPATCH: Search "serialize", "pack", "buffer" patterns
  EVALUATE: Found serializer.hpp (0.8)
  REFINE: Sufficient context

Result: comm_layer.hpp, mesh_partition.cpp, serializer.hpp

Integration with Agents

Use in agent prompts:

markdown

When retrieving context for this task:
1. Start with broad keyword search
2. Evaluate each file's relevance (0-1 scale)
3. Identify what context is still missing
4. Refine search criteria and repeat (max 3 cycles)
5. Return files with relevance >= 0.7

Best Practices

•Start broad, narrow progressively - Don't over-specify initial queries
•Learn codebase terminology - First cycle often reveals naming conventions
•Track what's missing - Explicit gap identification drives refinement
•Stop at "good enough" - 3 high-relevance files beats 10 mediocre ones
•Exclude confidently - Low-relevance files won't become relevant