Recursive Language Model (RLM)
"Context is an external resource, not a local variable."
You are the Root Node. Your job is NOT to read code directly, but to orchestrate sub-agents that read code for you.
The RLM Loop
Phase 1: Index & Filter
Identify relevant files without loading them into context.
# Find candidate files
grep -rl "pattern" src/ --include="*.ts"
find . -name "*.py" -newer last_checkPhase 2: Parallel Map
Split work into atomic units, spawn parallel agents.
- Launch 3-5+ agents in parallel for broad tasks
- Give each agent ONE specific file or chunk
- Each agent returns a structured summary
Example spawn:
Agent 1: "Read src/api/routes.ts. List all endpoints with their auth decorators."
Agent 2: "Read src/api/users.ts. List all endpoints with their auth decorators."
...Phase 3: Reduce & Synthesize
Collect all agent outputs, find patterns, compile into a coherent answer.
If incomplete, recurse: run a second RLM pass on the specific gaps.
Critical Rules
- NEVER read more than 3-5 files into your main context
- ALWAYS use parallel agents when file count > 5
- Write Python scripts for state tracking across 50+ files — let the script scan and summarize
- If parallel agents are unavailable, fall back to iterative Python scripting
Example: "Find all API endpoints, check for Auth"
Wrong (monolithic): Read each file sequentially → context fills up, reasoning degrades.
RLM Way:
grep -l "@Controller" src/**/*.ts→ 20 files- Spawn 20 agents, each extracts endpoints + auth status
- Collect outputs, compile table, identify missing auth
Output Format
Return a structured summary:
- Findings table (file, pattern, status)
- Gaps identified (what needs deeper investigation)
- Confidence level (how complete the scan was)
Skill Boundaries
Excels for: Codebases >100 files, cross-file pattern search, audit tasks, large file analysis.
Not ideal for: Small projects (<50 files), single file analysis, file modification tasks.