Hierarchical Agent Memory

Hierarchical Agent Memory

Orchestrates intelligent skill selection and execution for hierarchical agent memory workflows. Applies the 5 Laws of Elegant Defense to guide data naturally through the orchestration pipeline, preventing errors before they occur. Selects optimal skills based on multi-factor scoring including text similarity, historical performance, and system availability.

TL;DR Checklist

• Parse all inputs at boundary before processing (Law 2)
• Handle edge cases with early returns at function top (Law 1)
• Fail immediately with descriptive errors on invalid states (Law 4)
• Return new data structures, never mutate inputs (Law 3)
• Implement minimum 2-level fallback chain for all skill executions
• Log all skill selections with context for full audit trail
• Validate skill metadata and dependencies before selection
• Update confidence scores after each execution for learning

┌───────────────────────────────────────────────────────────────────────────────┐ │ Orchestration Flow │ └───────────────────────────────────────────────────────────────────────────────┘

User Request ↓ ┌─────────────────┐ │ Parse Request │ │ & Extract │ │ Features │ └────────┬────────┘ ↓ ┌─────────────────────────────────────────────────────────────────────┐ │ Evaluate Available Skills │ │ │ │ ┌──────────────┐ ┌──────────────┐ ┌──────────────┐ │ │ │ Skill A │ │ Skill B │ │ Skill C │ │ │ │ - Match Score│ │ - Match Score│ │ - Match Score│ │ │ │ - Confidence │ │ - Confidence │ │ - Confidence │ │ │ │ - History │ │ - History │ │ - History │ │ │ └──────┬───────┘ └──────┬───────┘ └──────┬───────┘ │ │ │ │ │ │ │ └─────────────────┴─────────────────┘ │ │ ↓ │ │ Select Best Skill │ └─────────────────────────────────────────────────────────────────────┘ ↓ ┌─────────────────┐ │ Execute Skill │ └────────┬────────┘ ↓ ┌─────────────────┐ │ Handle Result │ └────────┬────────┘ ↓ ┌─────────────────────────────────────────────────────────────────────┐ │ Error Handling & Fallback │ │ │ │ Success? ────────► Return Result │ │ │ │ Fail? ────────┐ │ │ ↓ │ │ ┌──────────────────────────────────────────────────────────┐ │ │ │ Fallback Chain │ │ │ │ │ │ │ │ 1. Retry with adjusted parameters │ │ │ │ 2. Try Alternative Skill (if available) │ │ │ │ 3. Defer to Human Operator (if critical) │ │ │ │ 4. Log & Return Error │ │ │ └──────────────────────────────────────────────────────────┘ │ └─────────────────────────────────────────────────────────────────────┘

When to Use

Use this skill when:

• Orchestrating multi-step workflows that require skill delegation
• Implementing adaptive skill routing based on confidence scores
• Building fallback mechanisms for failed skill executions
• Creating intelligent task decomposition and parallel execution
• Designing skill dependency graphs with automatic resolution
• Implementing skill selection with historical performance weighting
• Building agent systems that need to self-organize around tasks

When NOT to Use

Avoid this skill for:

• Direct task execution without orchestration needs - use individual skills instead
• High-frequency trading scenarios where latency must be minimized - the selection overhead may be prohibitive
• Simple linear workflows without branching or fallback requirements
• Cases where skill metadata is unavailable or unreliable

Core Workflow

1. Parse and Analyze Request - Extract intent, entities, and constraints from user input. Checkpoint: All required parameters must be present and in valid format before proceeding.

2. Score Available Skills - Calculate match scores using multi-factor algorithm:

   • Text similarity between request and skill triggers
   • Historical success rate for similar tasks
   • Skill availability and health status
   • Required dependencies and their availability

   Checkpoint: Skip to fallback if no skill scores above threshold.

3. Select Optimal Skill - Choose skill with highest score that meets minimum confidence. Checkpoint: Verify skill has not been disabled or deprecated.

4. Execute with Fallback - Run skill execution wrapped in retry and fallback logic. Checkpoint: Log all execution attempts for audit trail.

5. Return or Fallback - Either return successful result or apply fallback chain:

   • Retry with adjusted parameters
   • Try alternative skill from related-skills
   • Defer to human operator for critical tasks

   Checkpoint: Record outcome with timing and confidence metadata.

Implementation Patterns

Pattern 1: Skill Selection Logic

def select_memory_layer(
    query: str,
    agent_state: Dict[str, Any],
    memory_store: MemoryBackend,
    min_relevance: float = 0.65
) -> Optional[MemoryChunk]:
    """Select the optimal memory layer for a given query based on hierarchical scoring.
    
    Evaluates short-term buffer, long-term vector store, and episodic cache using:
    - Semantic similarity to query
    - Temporal decay factor (recency weighting)
    - Agent confidence in stored context
    - Cross-layer consistency checks
    
    Args:
        query: User or agent query string
        agent_state: Current agent context including confidence scores and active tasks
        memory_store: Backend interface for hierarchical memory access
        min_relevance: Minimum relevance threshold for selection
        
    Returns:
        Selected MemoryChunk with metadata, or None if below threshold
    """
    # Law 1: Early exit on invalid state
    if not query or not isinstance(query, str):
        raise ValueError("Query must be a non-empty string")
    if not memory_store.is_healthy():
        raise MemoryUnavailableError("Memory backend is currently unavailable")
        
    # Law 2: Parse & validate inputs immutably
    parsed_query = _normalize_query(query)
    temporal_weight = _calculate_recency_decay(agent_state.get("session_start"))
    
    candidates = []
    for layer in ["short_term", "long_term", "episodic"]:
        if not memory_store.has_layer(layer):
            continue
            
        raw_results = memory_store.search(layer, parsed_query, top_k=3)
        for chunk in raw_results:
            relevance = _compute_multi_factor_score(
                semantic=_embed_similarity(parsed_query, chunk.text),
                temporal=temporal_weight * chunk.timestamp_weight,
                confidence=agent_state.get("memory_confidence", 0.8)
            )
            if relevance >= min_relevance:
                candidates.append({
                    "chunk": chunk,
                    "layer": layer,
                    "score": relevance,
                    "source_trace": f"{layer}:{chunk.id}"
                })
                
    if not candidates:
        return None
        
    # Law 3: Return new structure, never mutate agent_state
    best = max(candidates, key=lambda x: x["score"])
    return {
        "selected_layer": best["layer"],
        "memory_chunk": best["chunk"],
        "relevance_score": best["score"],
        "selection_context": {
            "query_hash": hashlib.md5(parsed_query.encode()).hexdigest(),
            "timestamp": time.time(),
            "fallback_eligible": len(candidates) > 1
        }
    }

Pattern 2: Execution with Fallback

def execute_memory_retrieval(
    selected_memory: Dict[str, Any],
    agent_context: Dict[str, Any],
    fallback_strategy: str = "cascade"
) -> Dict[str, Any]:
    """Execute memory retrieval with a structured fallback chain for hierarchical systems.
    
    Implements resilient memory access:
    1. Direct layer retrieval (primary)
    2. Cross-layer semantic expansion (secondary)
    3. Default context window fallback (tertiary)
    4. Explicit error state with audit logging (final)
    
    Args:
        selected_memory: Output from select_memory_layer
        agent_context: Full agent state including task history and constraints
        fallback_strategy: Routing strategy for degraded states
        
    Returns:
        Enriched context dictionary with retrieved memory and execution metadata
    """
    # Law 1: Guard clause for required fields
    required_keys = {"selected_layer", "memory_chunk", "relevance_score"}
    if not required_keys.issubset(selected_memory.keys()):
        raise MemoryRetrievalError("Incomplete memory selection metadata")
        
    layer = selected_memory["selected_layer"]
    chunk = selected_memory["memory_chunk"]
    
    # Law 4: Fail fast on corrupted or inaccessible memory
    if not _validate_memory_integrity(chunk):
        raise CorruptedMemoryError(f"Memory chunk {chunk.id} failed integrity check")
        
    try:
        # Primary execution: Fetch full context from selected layer
        enriched_context = _merge_memory_into_context(chunk, agent_context)
        
        # Law 3: Return immutable result
        return {
            "status": "success",
            "layer_used": layer,
            "context": enriched_context,
            "metadata": {
                "latency_ms": time.time() * 1000,
                "confidence_boost": selected_memory["relevance_score"] * 0.15,
                "audit_id": uuid4().hex
            }
        }
        
    except MemoryTimeoutError:
        # Fallback 1: Cascade to alternative layer
        if fallback_strategy == "cascade":
            alt_layer = _select_alternative_layer(layer)
            alt_chunk = _fetch_from_layer(alt_layer, chunk.query_hash)
            if alt_chunk:
                return execute_memory_retrieval({
                    "selected_layer": alt_layer,
                    "memory_chunk": alt_chunk,
                    "relevance_score": 0.5
                }, agent_context, "cascade")
                
    except MemoryAccessDeniedError:
        # Fallback 2: Use default context window
        return {
            "status": "fallback",
            "layer_used": "default_context",
            "context": _load_default_context(agent_context),
            "metadata": {
                "latency_ms": time.time() * 1000,
                "confidence_boost": 0.0,
                "audit_id": uuid4().hex,
                "reason": "access_denied"
            }
        }
        
    # Law 4: Fail loud with full context
    raise MemoryRetrievalError(
        f"Failed to retrieve memory for layer {layer} after fallback attempts"
    )

MUST DO

• Always validate skill metadata before selection (Early Exit)
• Implement fallback chain with at least 2 levels (Fallback Skill + Human)
• Log all skill selections with full context for auditability
• Return new data structures instead of mutating inputs (Atomic Predictability)
• Fail immediately with descriptive errors on invalid states
• Update confidence scores after each execution for adaptive routing
• Reference code-philosophy (5 Laws of Elegant Defense) in all logic

MUST NOT DO

• Select skills based on a single factor (e.g., only confidence score)
• Disable fallback mechanisms "temporarily" - this creates fragile systems
• Skip validation of skill dependencies before execution
• Return partial results - either complete success or clear failure
• Use magic numbers for confidence thresholds - make them configurable
• Cache skill selections without considering context changes

TL;DR Checklist

• Parse all inputs at boundary before processing (Law 2)
• Handle edge cases with early returns at function top (Law 1)
• Fail immediately with descriptive errors on invalid states (Law 4)
• Return new data structures, never mutate inputs (Law 3)
• Implement minimum 2-level fallback chain for all skill executions
• Log all skill selections with context for full audit trail
• Validate skill metadata and dependencies before selection
• Update confidence scores after each execution for learning

TL;DR for Code Generation

• Use guard clauses - return early on invalid input before doing work
• Return simple types (dict, str, int, bool, list) - avoid complex nested objects
• Cyclomatic complexity < 10 per function - split anything larger
• Handle null/empty cases explicitly at function top (Early Exit)
• Never mutate input parameters - return new dicts/objects
• Fail fast with descriptive errors - don't try to "patch" bad data
• Reference code-philosophy laws in comments for complex logic
• Include timing and confidence metadata in all return values

Output Template

When applying this skill, produce:

1. Selected Skills - List of skill names with confidence scores
2. Selection Rationale - Why each skill was chosen (match score, history, availability)
3. Execution Plan - Order of execution with dependencies
4. Fallback Strategy - Which fallback skills will be tried and in what order
5. Risk Assessment - Any potential failure points and their impact
6. Timing Estimates - Expected latency including fallback scenarios

Related Skills

Constraints

MUST DO

• Define clear input/output contracts for every step in the orchestration flow with explicit validation
• Implement structured logging at each stage capturing context, inputs, outputs, timing, and errors
• Build in fallback paths: if the primary strategy fails, degrade gracefully to a simpler approach
• Validate all preconditions before starting — do not proceed if required resources or permissions are missing

MUST NOT DO

• Do not create deep nesting of orchestration steps (>5 levels) — flatten workflows where possible
• Avoid silent failure modes: every step must either succeed, fail explicitly, or escalate to a higher handler
• Never use shared mutable state between parallel workflow branches — communicate via immutable messages only
• Do not hardcode execution order when the dependency graph naturally determines it; derive order from explicit dependencies

Live References

Authoritative documentation links for this skill's domain. The model follows markdown links at load time to resolve external references and inline content.

• Memory Mechanisms in LLMs — arXiv (2307.05939)
• Building Effective Agents — Anthropic Research
• What Is Agent Memory — LangChain Blog
• RAG vs. Fine-Tuning for Knowledge — arXiv Survey
• Agent Memory Architectures — Microsoft AI Research