# Awesome Agentic Patterns A curated catalogue of AI agent design patterns. Every pattern is also available as Markdown at https://agentic-patterns.com/patterns/.md. Full text of all patterns: https://agentic-patterns.com/llms-full.txt. Usage notes for agents: https://agentic-patterns.com/agents. ## Patterns ### 402-first-machine-payments 402-First Machine Payments (Price-Before-Work Tool Purchases): Server returns an HTTP 402 price quote before doing work, and the agent pays the exact amount only if it fits its budget cap URL: https://agentic-patterns.com/patterns/402-first-machine-payments Markdown: https://agentic-patterns.com/patterns/402-first-machine-payments.md ### abstracted-code-representation-for-review Abstracted Code Representation for Review: Shows reviewers pseudocode, intent summaries, and logical diffs of code changes, with drill-down to the real code to confirm the mapping URL: https://agentic-patterns.com/patterns/abstracted-code-representation-for-review Markdown: https://agentic-patterns.com/patterns/abstracted-code-representation-for-review.md ### action-caching-replay Action Caching & Replay Pattern: Records each agent action with XPath and frame metadata so later runs replay it without LLM calls, with LLM fallback when replay fails URL: https://agentic-patterns.com/patterns/action-caching-replay Markdown: https://agentic-patterns.com/patterns/action-caching-replay.md ### action-selector-pattern Action-Selector Pattern: LLM maps user intent to a pre-approved action ID with schema-validated parameters, and tool outputs never return to the selector URL: https://agentic-patterns.com/patterns/action-selector-pattern Markdown: https://agentic-patterns.com/patterns/action-selector-pattern.md ### adaptive-sandbox-fanout-controller Adaptive Sandbox Fan-Out Controller: Starts a small batch of parallel sandboxes, then scales up, stops early, or refines the prompt based on early success, variance, and error signals URL: https://agentic-patterns.com/patterns/adaptive-sandbox-fanout-controller Markdown: https://agentic-patterns.com/patterns/adaptive-sandbox-fanout-controller.md ### agent-circuit-breaker Agent Circuit Breaker: Prevents agents from wasting tokens and time on repeatedly failing tools by tracking failure rates and temporarily disabling broken tool endpoints URL: https://agentic-patterns.com/patterns/agent-circuit-breaker Markdown: https://agentic-patterns.com/patterns/agent-circuit-breaker.md ### agent-modes-by-model-personality Agent Modes by Model Personality: Offers separate working modes, each with its own prompts, tools, UI, and expectations, tuned to one model's working style URL: https://agentic-patterns.com/patterns/agent-modes-by-model-personality Markdown: https://agentic-patterns.com/patterns/agent-modes-by-model-personality.md ### agent-reinforcement-fine-tuning Agent Reinforcement Fine-Tuning (Agent RFT): Trains model weights with reinforcement learning on real tool calls and custom graders to improve domain-specific tool use and multi-step reasoning URL: https://agentic-patterns.com/patterns/agent-reinforcement-fine-tuning Markdown: https://agentic-patterns.com/patterns/agent-reinforcement-fine-tuning.md ### agent-sdk-for-programmatic-control Agent SDK for Programmatic Control: Exposes agent functions through an SDK and CLI so code can run the agent headless with set tools, permissions, and resource limits URL: https://agentic-patterns.com/patterns/agent-sdk-for-programmatic-control Markdown: https://agentic-patterns.com/patterns/agent-sdk-for-programmatic-control.md ### agent-assisted-scaffolding Agent-Assisted Scaffolding: Agent generates initial files, boilerplate, and directory structure from a high-level description so developers start on core logic URL: https://agentic-patterns.com/patterns/agent-assisted-scaffolding Markdown: https://agentic-patterns.com/patterns/agent-assisted-scaffolding.md ### agent-driven-research Agent-Driven Research: Agent plans its own search queries, runs them across sources, reflects on gaps, and iterates until it can write a sourced report URL: https://agentic-patterns.com/patterns/agent-driven-research Markdown: https://agentic-patterns.com/patterns/agent-driven-research.md ### agent-first-tool-discovery Agent-First Tool Discovery: Build search indexes designed for agent consumers, returning structured tool metadata ranked by agent-relevant signals instead of human SEO metrics. URL: https://agentic-patterns.com/patterns/agent-first-tool-discovery Markdown: https://agentic-patterns.com/patterns/agent-first-tool-discovery.md ### agent-first-tooling-and-logging Agent-First Tooling and Logging: Designs tools and logs for agent consumption with one unified log stream, structured JSON output, and agent-aware CLI flags URL: https://agentic-patterns.com/patterns/agent-first-tooling-and-logging Markdown: https://agentic-patterns.com/patterns/agent-first-tooling-and-logging.md ### agent-friendly-workflow-design Agent-Friendly Workflow Design: Gives agents clear high-level goals, room for implementation choices, structured I/O, and plan review before execution URL: https://agentic-patterns.com/patterns/agent-friendly-workflow-design Markdown: https://agentic-patterns.com/patterns/agent-friendly-workflow-design.md ### agent-powered-codebase-qa-onboarding Agent-Powered Codebase Q&A / Onboarding: Agent indexes the codebase with embeddings and code graphs and answers natural-language questions about where code is and how it behaves URL: https://agentic-patterns.com/patterns/agent-powered-codebase-qa-onboarding Markdown: https://agentic-patterns.com/patterns/agent-powered-codebase-qa-onboarding.md ### agentic-search-over-vector-embeddings Agentic Search Over Vector Embeddings: Replaces vector indexes with agent-driven grep, find, and file traversal that searches current file state on demand and refines iteratively URL: https://agentic-patterns.com/patterns/agentic-search-over-vector-embeddings Markdown: https://agentic-patterns.com/patterns/agentic-search-over-vector-embeddings.md ### ai-web-search-agent-loop AI Web Search Agent Loop: Coordinator agent translates queries, spawns parallel search workers across domains and time ranges, refines iteratively, and answers with citations URL: https://agentic-patterns.com/patterns/ai-web-search-agent-loop Markdown: https://agentic-patterns.com/patterns/ai-web-search-agent-loop.md ### ai-accelerated-learning-and-skill-development AI-Accelerated Learning and Skill Development: Uses AI assistants as tutors that explain errors and concepts, offer alternatives, and fade support as the developer gains skill URL: https://agentic-patterns.com/patterns/ai-accelerated-learning-and-skill-development Markdown: https://agentic-patterns.com/patterns/ai-accelerated-learning-and-skill-development.md ### ai-assisted-code-review-verification AI-Assisted Code Review / Verification: Uses AI tools to flag issues, summarize change intent, and explain code so human reviewers focus on intent and business logic URL: https://agentic-patterns.com/patterns/ai-assisted-code-review-verification Markdown: https://agentic-patterns.com/patterns/ai-assisted-code-review-verification.md ### anti-reward-hacking-grader-design Anti-Reward-Hacking Grader Design: Design reward functions with multi-criteria evaluation and iterative hardening to prevent models from gaming graders, ensuring training rewards align with actual task quality. URL: https://agentic-patterns.com/patterns/anti-reward-hacking-grader-design Markdown: https://agentic-patterns.com/patterns/anti-reward-hacking-grader-design.md ### approval-bound-transactional-write Approval-Bound Transactional Write: Binds a human approval to one exact plan version, then applies the write, checks the postcondition and records the idempotency key in a single database transaction URL: https://agentic-patterns.com/patterns/approval-bound-transactional-write Markdown: https://agentic-patterns.com/patterns/approval-bound-transactional-write.md ### multi-step-analysis-pipeline-orchestration Artifact-Driven Analysis Pipeline Orchestration: Has an agent run independent analysis scripts, read their structured reports, and merge them into one final report or visualization URL: https://agentic-patterns.com/patterns/multi-step-analysis-pipeline-orchestration Markdown: https://agentic-patterns.com/patterns/multi-step-analysis-pipeline-orchestration.md ### asynchronous-coding-agent-pipeline Asynchronous Coding Agent Pipeline: Splits inference, tool execution, reward modeling, and learning into asynchronous workers linked by message queues so GPUs stay busy URL: https://agentic-patterns.com/patterns/asynchronous-coding-agent-pipeline Markdown: https://agentic-patterns.com/patterns/asynchronous-coding-agent-pipeline.md ### authenticated-authority-channel Authenticated Authority Channel: Preserve a distinguishable channel for authenticated intent so retrieved content can inform reasoning without granting itself authority. URL: https://agentic-patterns.com/patterns/authenticated-authority-channel Markdown: https://agentic-patterns.com/patterns/authenticated-authority-channel.md ### autonomous-workflow-agent-architecture Autonomous Workflow Agent Architecture: Runs multi-step engineering workflows in containers with tmux sessions, sleep-based monitoring, checkpoints, and context-aware error recovery URL: https://agentic-patterns.com/patterns/autonomous-workflow-agent-architecture Markdown: https://agentic-patterns.com/patterns/autonomous-workflow-agent-architecture.md ### background-agent-ci Background Agent with CI Feedback: Runs the agent in the background on its own branch and uses CI results as feedback to patch failures until green or blocked URL: https://agentic-patterns.com/patterns/background-agent-ci Markdown: https://agentic-patterns.com/patterns/background-agent-ci.md ### black-box-skill-invocation Black-Box Skill Invocation: Shares skills through input and output schemas only and runs them on the provider side, so prompts and code never cross the boundary URL: https://agentic-patterns.com/patterns/black-box-skill-invocation Markdown: https://agentic-patterns.com/patterns/black-box-skill-invocation.md ### board-mediated-inter-agent-coordination Board-Mediated Async Inter-Agent Coordination: Routes agent-to-agent messages through comment threads on board cards, with a prefix that creates an inbox card to notify the target agent URL: https://agentic-patterns.com/patterns/board-mediated-inter-agent-coordination Markdown: https://agentic-patterns.com/patterns/board-mediated-inter-agent-coordination.md ### budget-aware-model-routing-with-hard-cost-caps Budget-Aware Model Routing with Hard Cost Caps: Routes each request to the cheapest model that meets its needs under hard cost caps, and escalates only when quality gates fail URL: https://agentic-patterns.com/patterns/budget-aware-model-routing-with-hard-cost-caps Markdown: https://agentic-patterns.com/patterns/budget-aware-model-routing-with-hard-cost-caps.md ### burn-the-boats Burn the Boats: Removes working but obsolete features on a hard, announced deadline so the team and users move to the new approach URL: https://agentic-patterns.com/patterns/burn-the-boats Markdown: https://agentic-patterns.com/patterns/burn-the-boats.md ### canary-rollout-and-automatic-rollback-for-agent-policy-changes Canary Rollout and Automatic Rollback for Agent Policy Changes: Ships agent policy changes to a small traffic slice first, monitors guardrail metrics, and rolls back to the last stable version automatically URL: https://agentic-patterns.com/patterns/canary-rollout-and-automatic-rollback-for-agent-policy-changes Markdown: https://agentic-patterns.com/patterns/canary-rollout-and-automatic-rollback-for-agent-policy-changes.md ### capability-escrow-receipt Capability-Escrow-Receipt: Binds a signed capability listing, an atomic escrow-plus-hire step, and a signed work receipt into one loop for agent-to-agent payment URL: https://agentic-patterns.com/patterns/capability-escrow-receipt Markdown: https://agentic-patterns.com/patterns/capability-escrow-receipt.md ### chain-of-thought-monitoring-interruption Chain-of-Thought Monitoring & Interruption: Streams agent reasoning and tool calls in real time so a human can interrupt and redirect early when the approach is wrong URL: https://agentic-patterns.com/patterns/chain-of-thought-monitoring-interruption Markdown: https://agentic-patterns.com/patterns/chain-of-thought-monitoring-interruption.md ### classify-then-act-for-background-agents Classify-Then-Act for Background Agents: Sorts each task into has-default, needs-intent, or out-of-scope, builds only has-default work in a sandbox, and queues it for human review URL: https://agentic-patterns.com/patterns/classify-then-act-for-background-agents Markdown: https://agentic-patterns.com/patterns/classify-then-act-for-background-agents.md ### cli-first-skill-design CLI-First Skill Design: Builds each skill as a standalone CLI with JSON output and exit codes so humans and agents use the same interface URL: https://agentic-patterns.com/patterns/cli-first-skill-design Markdown: https://agentic-patterns.com/patterns/cli-first-skill-design.md ### cli-native-agent-orchestration CLI-Native Agent Orchestration: Drives agents through their non-interactive CLI mode (claude -p with JSON output and exit codes) so Makefiles, Git hooks, cron jobs, and CI can script and replay agent runs URL: https://agentic-patterns.com/patterns/cli-native-agent-orchestration Markdown: https://agentic-patterns.com/patterns/cli-native-agent-orchestration.md ### code-first-tool-interface-pattern Code Mode MCP Tool Interface Improvement Pattern: LLMs generate TypeScript code to orchestrate MCP tools in ephemeral V8 isolates, eliminating token-heavy round-trips and enabling efficient multi-step workflows with large measured token savings. URL: https://agentic-patterns.com/patterns/code-first-tool-interface-pattern Markdown: https://agentic-patterns.com/patterns/code-first-tool-interface-pattern.md ### code-over-api-pattern Code-Over-API Pattern: Agent writes code that calls tools and filters data inside a sandbox, so only summaries and samples return to the context window URL: https://agentic-patterns.com/patterns/code-over-api-pattern Markdown: https://agentic-patterns.com/patterns/code-over-api-pattern.md ### code-then-execute-pattern Code-Then-Execute Pattern: LLM writes a sandboxed program or DSL script, a static taint checker verifies data flows, and an interpreter runs it in a locked sandbox URL: https://agentic-patterns.com/patterns/code-then-execute-pattern Markdown: https://agentic-patterns.com/patterns/code-then-execute-pattern.md ### codebase-optimization-for-agents Codebase Optimization for Agents: Optimizes tooling, CLIs, tests, and docs for agents first, with one-command verify loops and machine-readable output, even if human DX regresses URL: https://agentic-patterns.com/patterns/codebase-optimization-for-agents Markdown: https://agentic-patterns.com/patterns/codebase-optimization-for-agents.md ### commitment-ledger-reality-gated-credit Commitment Ledger with Reality-Gated Credit: Records agent promises and credits retrieved memory only after an externally verifiable outcome settles. URL: https://agentic-patterns.com/patterns/commitment-ledger-reality-gated-credit Markdown: https://agentic-patterns.com/patterns/commitment-ledger-reality-gated-credit.md ### compounding-engineering-pattern Compounding Engineering Pattern: After each feature, codifies agent mistakes and learnings into CLAUDE.md, slash commands, subagents, and hooks so the next feature is easier to build URL: https://agentic-patterns.com/patterns/compounding-engineering-pattern Markdown: https://agentic-patterns.com/patterns/compounding-engineering-pattern.md ### parallel-tool-execution Conditional Parallel Tool Execution: Runs a batch of tool calls in parallel when all are read-only and in sequence when any modifies state, then returns results in request order URL: https://agentic-patterns.com/patterns/parallel-tool-execution Markdown: https://agentic-patterns.com/patterns/parallel-tool-execution.md ### consequence-family-coverage-audit Consequence-Family Coverage Audit: Audit an agent's risk policy by enumerating families of consequence and asking which rule covers each, because a hand-written risk list reliably encodes one family and stays silent on the rest URL: https://agentic-patterns.com/patterns/consequence-family-coverage-audit Markdown: https://agentic-patterns.com/patterns/consequence-family-coverage-audit.md ### context-budget-as-a-governed-resource Context Budget as a Governed Resource: Measures always-loaded context per source, alerts on trend growth, and puts hard spend caps and fan-out bounds on unattended agent runs URL: https://agentic-patterns.com/patterns/context-budget-as-a-governed-resource Markdown: https://agentic-patterns.com/patterns/context-budget-as-a-governed-resource.md ### context-window-anxiety-management Context Window Anxiety Management: Enables a large context window but caps usage lower, and adds prompts that state the token budget so the model does not rush to finish URL: https://agentic-patterns.com/patterns/context-window-anxiety-management Markdown: https://agentic-patterns.com/patterns/context-window-anxiety-management.md ### context-window-auto-compaction Context Window Auto-Compaction: Catches context overflow errors, compacts and validates the transcript with capped reserve headroom, and retries the request automatically URL: https://agentic-patterns.com/patterns/context-window-auto-compaction Markdown: https://agentic-patterns.com/patterns/context-window-auto-compaction.md ### context-minimization-pattern Context-Minimization Pattern: Removes untrusted user text and tool output from context after it becomes a safe structured artifact, so later steps see only trusted data URL: https://agentic-patterns.com/patterns/context-minimization-pattern Markdown: https://agentic-patterns.com/patterns/context-minimization-pattern.md ### continuous-autonomous-task-loop-pattern Continuous Autonomous Task Loop Pattern: Runs a loop where subagents pick the next task from a todo file, execute it in fresh context, commit it, and back off on rate limits URL: https://agentic-patterns.com/patterns/continuous-autonomous-task-loop-pattern Markdown: https://agentic-patterns.com/patterns/continuous-autonomous-task-loop-pattern.md ### criticgpt-style-evaluation CriticGPT-Style Code Review: Runs a specialized critic model over generated code to find bugs, security flaws, and quality issues, and feeds critical findings back for regeneration URL: https://agentic-patterns.com/patterns/criticgpt-style-evaluation Markdown: https://agentic-patterns.com/patterns/criticgpt-style-evaluation.md ### cross-agent-lesson-sharing Cross-Agent Lesson Sharing via Git: Agents write solved problems as markdown lessons in a shared Git repo and search them before debugging, with GitHub Issues as the coordination layer URL: https://agentic-patterns.com/patterns/cross-agent-lesson-sharing Markdown: https://agentic-patterns.com/patterns/cross-agent-lesson-sharing.md ### cross-cycle-consensus-relay Cross-Cycle Consensus Relay: Each loop cycle reads a structured relay document, does its work, and atomically writes back decisions, open questions, and the single next action URL: https://agentic-patterns.com/patterns/cross-cycle-consensus-relay Markdown: https://agentic-patterns.com/patterns/cross-cycle-consensus-relay.md ### cross-domain-agent-conflict-resolution Cross-Domain Agent Conflict Resolution: A coordination layer that cross-references recommendations from independent domain agents, detects conflicts on shared resources, and resolves them through policy-as-code. URL: https://agentic-patterns.com/patterns/cross-domain-agent-conflict-resolution Markdown: https://agentic-patterns.com/patterns/cross-domain-agent-conflict-resolution.md ### cross-protocol-agent-discovery Cross-Protocol Agent Discovery: Aggregates agent metadata across fragmented registries and protocols into a unified, protocol-agnostic discovery layer (proposed; no neutral implementation verified) URL: https://agentic-patterns.com/patterns/cross-protocol-agent-discovery Markdown: https://agentic-patterns.com/patterns/cross-protocol-agent-discovery.md ### cryptographic-governance-audit-trail Cryptographic Governance Audit Trail: Middleware checks each tool call against a policy file, then signs a receipt of the action with ML-DSA and appends it to a tamper-evident chain URL: https://agentic-patterns.com/patterns/cryptographic-governance-audit-trail Markdown: https://agentic-patterns.com/patterns/cryptographic-governance-audit-trail.md ### curated-code-context-window Curated Code Context Window: A search subagent finds the few relevant files for a task and injects only top snippets or summaries into the main coding agent's context URL: https://agentic-patterns.com/patterns/curated-code-context-window Markdown: https://agentic-patterns.com/patterns/curated-code-context-window.md ### custom-sandboxed-background-agent Custom Sandboxed Background Agent: Builds an in-house background coding agent that runs in sandboxed dev environments, streams progress over WebSocket, and iterates on compiler and test feedback URL: https://agentic-patterns.com/patterns/custom-sandboxed-background-agent Markdown: https://agentic-patterns.com/patterns/custom-sandboxed-background-agent.md ### dead-mans-switch-for-scheduled-agent-jobs Dead-Man's Switch for Scheduled Agent Jobs: Each scheduled job writes a success sentinel to its log, and a separate checker files a task when the sentinel is missing or stale URL: https://agentic-patterns.com/patterns/dead-mans-switch-for-scheduled-agent-jobs Markdown: https://agentic-patterns.com/patterns/dead-mans-switch-for-scheduled-agent-jobs.md ### declarative-multi-agent-topology-definition Declarative Multi-Agent Topology Definition: Define multi-agent systems declaratively in a single topology file — agents, flows, gates, hooks, group chats — then compile to platform-specific configurations for any agentic framework. URL: https://agentic-patterns.com/patterns/declarative-multi-agent-topology-definition Markdown: https://agentic-patterns.com/patterns/declarative-multi-agent-topology-definition.md ### democratization-of-tooling-via-agents Democratization of Tooling via Agents: Non-programmers describe a tool in natural language and iterate with an AI agent that generates and fixes the code for dashboards, scripts, or small apps URL: https://agentic-patterns.com/patterns/democratization-of-tooling-via-agents Markdown: https://agentic-patterns.com/patterns/democratization-of-tooling-via-agents.md ### denial-tracking-permission-escalation Denial Tracking & Permission Escalation: Track repeated tool denials and auto-escalate to blanket permission prompts or fallback strategies, preventing wasted iterations in non-interactive contexts. URL: https://agentic-patterns.com/patterns/denial-tracking-permission-escalation Markdown: https://agentic-patterns.com/patterns/denial-tracking-permission-escalation.md ### design-time-file-partition Design-Time File Partition as Concurrency Control: Each task declares every path it will write; overlapping tasks get a blocking dependency so only disjoint work runs at once on a shared checkout URL: https://agentic-patterns.com/patterns/design-time-file-partition Markdown: https://agentic-patterns.com/patterns/design-time-file-partition.md ### deterministic-grader-in-the-loop Deterministic Grader in the Loop: Replace the LLM-as-judge in an agent's self-review loop with a deterministic scorer that returns the raw counts behind each sub-score, so the agent has something specific to fix. URL: https://agentic-patterns.com/patterns/deterministic-grader-in-the-loop Markdown: https://agentic-patterns.com/patterns/deterministic-grader-in-the-loop.md ### deterministic-security-scanning-build-loop Deterministic Security Scanning Build Loop: Adds SAST, SCA, and secret scanners to the build target that the agent must run after each change, so scanner failures force the agent to fix the code URL: https://agentic-patterns.com/patterns/deterministic-security-scanning-build-loop Markdown: https://agentic-patterns.com/patterns/deterministic-security-scanning-build-loop.md ### deterministic-threat-rule-scanning Deterministic Threat Rule Scanning: Apply deterministic regex rules as a first-pass security layer to detect known threat patterns in AI agent tool calls and skill definitions. URL: https://agentic-patterns.com/patterns/deterministic-threat-rule-scanning Markdown: https://agentic-patterns.com/patterns/deterministic-threat-rule-scanning.md ### deterministic-zero-llm-orchestration Deterministic Zero-LLM Orchestration: A deterministic code orchestrator splits goals, runs parallel coding agents, verifies with tests, and commits, spending LLM tokens only on the one-time goal decomposition URL: https://agentic-patterns.com/patterns/deterministic-zero-llm-orchestration Markdown: https://agentic-patterns.com/patterns/deterministic-zero-llm-orchestration.md ### dev-tooling-assumptions-reset Dev Tooling Assumptions Reset: Audits dev tools for human-effort assumptions and replaces tickets, PR ceremony, and sprints with immediate agent dispatch, variations, and automated tests URL: https://agentic-patterns.com/patterns/dev-tooling-assumptions-reset Markdown: https://agentic-patterns.com/patterns/dev-tooling-assumptions-reset.md ### discrete-phase-separation Discrete Phase Separation: Splits work into separate research, planning, and implementation conversations, each with fresh context, passing only distilled outputs between phases URL: https://agentic-patterns.com/patterns/discrete-phase-separation Markdown: https://agentic-patterns.com/patterns/discrete-phase-separation.md ### disposable-scaffolding-over-durable-features Disposable Scaffolding Over Durable Features: Treats code around the model as temporary scaffolding: build the simplest thing that works now, mark it disposable, and remove it when models improve URL: https://agentic-patterns.com/patterns/disposable-scaffolding-over-durable-features Markdown: https://agentic-patterns.com/patterns/disposable-scaffolding-over-durable-features.md ### distributed-execution-cloud-workers Distributed Execution with Cloud Workers: Runs many agent sessions in parallel on cloud workers, each in its own git worktree, with dependency-aware scheduling, merge coordination, and approval gates URL: https://agentic-patterns.com/patterns/distributed-execution-cloud-workers Markdown: https://agentic-patterns.com/patterns/distributed-execution-cloud-workers.md ### dogfooding-with-rapid-iteration-for-agent-improvement Dogfooding with Rapid Iteration for Agent Improvement: The agent team uses its own agent for daily work, collects feedback in low-friction channels, and ships features internally first to validate or discard them URL: https://agentic-patterns.com/patterns/dogfooding-with-rapid-iteration-for-agent-improvement Markdown: https://agentic-patterns.com/patterns/dogfooding-with-rapid-iteration-for-agent-improvement.md ### dual-llm-pattern Dual LLM Pattern: Splits work between a privileged LLM that calls tools but never sees untrusted data and a quarantined LLM that reads untrusted data but has no tools URL: https://agentic-patterns.com/patterns/dual-llm-pattern Markdown: https://agentic-patterns.com/patterns/dual-llm-pattern.md ### dual-rail-message-delivery Dual-Rail Message Delivery: Sends every inter-agent message on two independent channels through one entry point, alerts when the rails diverge, and requires ACKs to confirm completion URL: https://agentic-patterns.com/patterns/dual-rail-message-delivery Markdown: https://agentic-patterns.com/patterns/dual-rail-message-delivery.md ### dual-use-tool-design Dual-Use Tool Design: Builds each tool with one interface and implementation that both humans and agents can call, with the same output, permissions, and logs URL: https://agentic-patterns.com/patterns/dual-use-tool-design Markdown: https://agentic-patterns.com/patterns/dual-use-tool-design.md ### dynamic-context-injection Dynamic Context Injection: Lets users inject files, folders, or saved prompts into the agent's context mid-session with @-mentions and custom slash commands URL: https://agentic-patterns.com/patterns/dynamic-context-injection Markdown: https://agentic-patterns.com/patterns/dynamic-context-injection.md ### egress-lockdown-no-exfiltration-channel Egress Lockdown (No-Exfiltration Channel): Puts the agent behind a default-deny egress firewall with allowlisted destinations so stolen data has no outbound channel URL: https://agentic-patterns.com/patterns/egress-lockdown-no-exfiltration-channel Markdown: https://agentic-patterns.com/patterns/egress-lockdown-no-exfiltration-channel.md ### episodic-memory-retrieval-injection Episodic Memory Retrieval & Injection: Writes a structured memory record after each episode and injects the top-k similar past memories as hints when a new task starts URL: https://agentic-patterns.com/patterns/episodic-memory-retrieval-injection Markdown: https://agentic-patterns.com/patterns/episodic-memory-retrieval-injection.md ### evidence-questions-not-verdict-questions Evidence Questions, Not Verdict Questions: Never ask the model for allow/ask/block. Ask it narrow, typed, observational questions and let code reduce the answers to the verdict, because the collapsed question is the low-confidence one. URL: https://agentic-patterns.com/patterns/evidence-questions-not-verdict-questions Markdown: https://agentic-patterns.com/patterns/evidence-questions-not-verdict-questions.md ### evidence-layered-evaluation-for-interactive-agents Evidence-Layered Evaluation for Interactive Agents: Scores interactive agent runs on a task outcome assertion and links each result to layered evidence: actions, screenshots, replays, network, and messages URL: https://agentic-patterns.com/patterns/evidence-layered-evaluation-for-interactive-agents Markdown: https://agentic-patterns.com/patterns/evidence-layered-evaluation-for-interactive-agents.md ### exact-action-authorization-binding Exact-Action Authorization Binding: Bind a short-lived approval to the complete action presented to the reviewer, then rederive and compare that identity at the acting boundary before execution. URL: https://agentic-patterns.com/patterns/exact-action-authorization-binding Markdown: https://agentic-patterns.com/patterns/exact-action-authorization-binding.md ### explicit-posterior-sampling-planner Explicit Posterior-Sampling Planner: Embeds Posterior Sampling for RL in the LLM's reasoning: sample a task model, plan, act, observe reward, and update the posterior URL: https://agentic-patterns.com/patterns/explicit-posterior-sampling-planner Markdown: https://agentic-patterns.com/patterns/explicit-posterior-sampling-planner.md ### extended-coherence-work-sessions Extended Coherence Work Sessions: Combines models with long coherence windows with context compaction, prompt caching, and persisted state so agents stay on task for hours URL: https://agentic-patterns.com/patterns/extended-coherence-work-sessions Markdown: https://agentic-patterns.com/patterns/extended-coherence-work-sessions.md ### external-credential-sync External Credential Sync: Reads AI credentials from other tools' keychains and config files and syncs them into the agent's store, with TTL-cached reads, refresh fences, and persisted ownership URL: https://agentic-patterns.com/patterns/external-credential-sync Markdown: https://agentic-patterns.com/patterns/external-credential-sync.md ### factory-over-assistant Factory over Assistant: Spawns several autonomous agents in parallel with automated feedback loops such as tests and builds, and checks on them later instead of watching one agent URL: https://agentic-patterns.com/patterns/factory-over-assistant Markdown: https://agentic-patterns.com/patterns/factory-over-assistant.md ### failover-aware-model-fallback Failover-Aware Model Fallback: Classifies each model failure by reason and falls back along a model chain only for retryable errors, failing fast on auth, billing, and user aborts URL: https://agentic-patterns.com/patterns/failover-aware-model-fallback Markdown: https://agentic-patterns.com/patterns/failover-aware-model-fallback.md ### fair-share-task-queues-with-heartbeat-leases Fair-Share Task Queues with Heartbeat Leases: Distributes tasks across a shared agent fleet using weighted deficit-round-robin scheduling, with heartbeat leases and dead-letter retries to recover from crashed workers URL: https://agentic-patterns.com/patterns/fair-share-task-queues-with-heartbeat-leases Markdown: https://agentic-patterns.com/patterns/fair-share-task-queues-with-heartbeat-leases.md ### feature-list-as-immutable-contract Feature List as Immutable Contract: Defines every feature up front in a JSON list with acceptance steps; the agent may only flip a feature to passing after it verifies it URL: https://agentic-patterns.com/patterns/feature-list-as-immutable-contract Markdown: https://agentic-patterns.com/patterns/feature-list-as-immutable-contract.md ### filesystem-based-agent-state Filesystem-Based Agent State: Agents persist intermediate results and working state to files, creating durable checkpoints that enable workflow resumption, recovery from failures, and support for long-running tasks. URL: https://agentic-patterns.com/patterns/filesystem-based-agent-state Markdown: https://agentic-patterns.com/patterns/filesystem-based-agent-state.md ### filesystem-mediated-host-delegation Filesystem-Mediated Host Delegation: Lets a sandboxed agent write request files to a shared spool directory that a host daemon runs against a whitelist, with idempotent results the agent polls for URL: https://agentic-patterns.com/patterns/filesystem-mediated-host-delegation Markdown: https://agentic-patterns.com/patterns/filesystem-mediated-host-delegation.md ### frontier-focused-development Frontier-Focused Development: Targets only state-of-the-art models, picks the best model per use case with no user model selector, and expects to rework the product every few months URL: https://agentic-patterns.com/patterns/frontier-focused-development Markdown: https://agentic-patterns.com/patterns/frontier-focused-development.md ### graph-of-thoughts Graph of Thoughts (GoT): Represents reasoning as a directed graph of thoughts so the model can branch, aggregate, refine, and revisit reasoning paths URL: https://agentic-patterns.com/patterns/graph-of-thoughts Markdown: https://agentic-patterns.com/patterns/graph-of-thoughts.md ### hook-based-safety-guard-rails Hook-Based Safety Guard Rails for Autonomous Code Agents: Runs small shell scripts on PreToolUse and PostToolUse hooks to block destructive commands, lint edits, and warn on context use outside the agent's reasoning URL: https://agentic-patterns.com/patterns/hook-based-safety-guard-rails Markdown: https://agentic-patterns.com/patterns/hook-based-safety-guard-rails.md ### human-in-loop-approval-framework Human-in-the-Loop Approval Framework: Systematically insert human approval gates for designated high-risk functions while maintaining agent autonomy for safe operations, with multi-channel approval interfaces and comprehensive audit trails. URL: https://agentic-patterns.com/patterns/human-in-loop-approval-framework Markdown: https://agentic-patterns.com/patterns/human-in-loop-approval-framework.md ### hybrid-llm-code-workflow-coordinator Hybrid LLM/Code Workflow Coordinator: Lets each workflow pick an LLM or a code script as its coordinator, so teams prototype with the LLM and move to reviewed code when determinism matters URL: https://agentic-patterns.com/patterns/hybrid-llm-code-workflow-coordinator Markdown: https://agentic-patterns.com/patterns/hybrid-llm-code-workflow-coordinator.md ### incident-to-eval-synthesis Incident-to-Eval Synthesis: Converts each production incident into executable eval cases with pass/fail criteria and gates future releases on them URL: https://agentic-patterns.com/patterns/incident-to-eval-synthesis Markdown: https://agentic-patterns.com/patterns/incident-to-eval-synthesis.md ### inference-healed-code-review-reward Inference-Healed Code Review Reward: Replaces a binary tests-passed reward with a critic that scores correctness, style, performance, and security, explains low scores, and combines them URL: https://agentic-patterns.com/patterns/inference-healed-code-review-reward Markdown: https://agentic-patterns.com/patterns/inference-healed-code-review-reward.md ### inference-time-scaling Inference-Time Scaling: Spends extra compute at inference time with best-of-N sampling, longer reasoning, self-refinement, search, and verification to improve output quality URL: https://agentic-patterns.com/patterns/inference-time-scaling Markdown: https://agentic-patterns.com/patterns/inference-time-scaling.md ### initializer-maintainer-dual-agent Initializer-Maintainer Dual Agent Architecture: Uses a one-time initializer agent to create the feature list, progress files, and bootstrap script, then a coding agent that resumes from them one feature per session URL: https://agentic-patterns.com/patterns/initializer-maintainer-dual-agent Markdown: https://agentic-patterns.com/patterns/initializer-maintainer-dual-agent.md ### intelligent-bash-tool-execution Intelligent Bash Tool Execution: Runs agent shell commands through a multi-mode executor that uses a PTY when needed, falls back to direct exec, and manages approvals, background jobs, and signals URL: https://agentic-patterns.com/patterns/intelligent-bash-tool-execution Markdown: https://agentic-patterns.com/patterns/intelligent-bash-tool-execution.md ### inversion-of-control Inversion of Control: Gives the agent tools, a high-level objective, and guardrails, then lets it choose sequencing and recovery while humans set policy and review URL: https://agentic-patterns.com/patterns/inversion-of-control Markdown: https://agentic-patterns.com/patterns/inversion-of-control.md ### isolated-vm-per-rl-rollout Isolated VM per RL Rollout: Spin up an isolated virtual machine for each RL rollout to prevent cross-contamination between parallel agent executions, ensuring safe training with destructive tool access. URL: https://agentic-patterns.com/patterns/isolated-vm-per-rl-rollout Markdown: https://agentic-patterns.com/patterns/isolated-vm-per-rl-rollout.md ### iterative-multi-agent-brainstorming Iterative Multi-Agent Brainstorming: Spawns several agents in parallel on the same problem, often with different perspectives, then merges their ideas into one set of options URL: https://agentic-patterns.com/patterns/iterative-multi-agent-brainstorming Markdown: https://agentic-patterns.com/patterns/iterative-multi-agent-brainstorming.md ### iterative-prompt-skill-refinement Iterative Prompt & Skill Refinement: Combines a feedback channel, editable prompt documents, log-driven skill fixes, and usage dashboards to improve agent prompts and skills continuously URL: https://agentic-patterns.com/patterns/iterative-prompt-skill-refinement Markdown: https://agentic-patterns.com/patterns/iterative-prompt-skill-refinement.md ### judge-gated-finding-exchange-between-sub-agents Judge-Gated Finding Exchange Between Parallel Sub-Agents: Puts an append-only board between parallel sub-agents and lets a small judge decide what goes on it, who receives each note, and when a later note retracts an earlier one URL: https://agentic-patterns.com/patterns/judge-gated-finding-exchange-between-sub-agents Markdown: https://agentic-patterns.com/patterns/judge-gated-finding-exchange-between-sub-agents.md ### lane-based-execution-queueing Lane-Based Execution Queueing: Routes work into named lanes, each with its own queue and concurrency limit, so sessions never interleave and background jobs never block users URL: https://agentic-patterns.com/patterns/lane-based-execution-queueing Markdown: https://agentic-patterns.com/patterns/lane-based-execution-queueing.md ### language-agent-tree-search-lats Language Agent Tree Search (LATS): Runs Monte Carlo Tree Search over reasoning steps, with the LLM generating candidate actions and scoring partial solutions to pick the best path URL: https://agentic-patterns.com/patterns/language-agent-tree-search-lats Markdown: https://agentic-patterns.com/patterns/language-agent-tree-search-lats.md ### latent-demand-product-discovery Latent Demand Product Discovery: Builds hackable, extensible products, watches how power users repurpose them, and turns the most frequent workarounds into supported features URL: https://agentic-patterns.com/patterns/latent-demand-product-discovery Markdown: https://agentic-patterns.com/patterns/latent-demand-product-discovery.md ### layered-configuration-context Layered Configuration Context: Loads context files from enterprise, user, project, and local levels automatically and merges them into the agent's baseline context URL: https://agentic-patterns.com/patterns/layered-configuration-context Markdown: https://agentic-patterns.com/patterns/layered-configuration-context.md ### lethal-trifecta-threat-model Lethal Trifecta Threat Model: Classifies each tool by private-data access, untrusted-content exposure, and external communication, and blocks any execution path that has all three URL: https://agentic-patterns.com/patterns/lethal-trifecta-threat-model Markdown: https://agentic-patterns.com/patterns/lethal-trifecta-threat-model.md ### llm-map-reduce-pattern LLM Map-Reduce Pattern: Processes each untrusted document in its own sandboxed LLM call with a constrained output, then aggregates the results with deterministic code URL: https://agentic-patterns.com/patterns/llm-map-reduce-pattern Markdown: https://agentic-patterns.com/patterns/llm-map-reduce-pattern.md ### llm-observability LLM Observability: Sends agent runs to an LLM observability platform for span-level traces of each LLM call and tool use, plus aggregate cost, latency, and success metrics URL: https://agentic-patterns.com/patterns/llm-observability Markdown: https://agentic-patterns.com/patterns/llm-observability.md ### llm-friendly-api-design LLM-Friendly API Design: Designs APIs with visible versions, self-descriptive names and schemas, simple calls, actionable errors, and few indirection levels so LLMs call them correctly URL: https://agentic-patterns.com/patterns/llm-friendly-api-design Markdown: https://agentic-patterns.com/patterns/llm-friendly-api-design.md ### local-first-credential-broker Local-First Credential Broker: Keep raw secrets out of the agent process by injecting credentials at the network layer through a local broker, rather than handing the agent environment variables or config files. URL: https://agentic-patterns.com/patterns/local-first-credential-broker Markdown: https://agentic-patterns.com/patterns/local-first-credential-broker.md ### markdown-polis-coordination Markdown Polis — Multi-Vendor Agent Coordination via Filesystem Constitution: Coordinates agents from different vendors through versioned markdown files: capability cards, work contracts, and bandit routing that learns from settled work URL: https://agentic-patterns.com/patterns/markdown-polis-coordination Markdown: https://agentic-patterns.com/patterns/markdown-polis-coordination.md ### mcp-pattern-injection MCP Pattern Injection: Runs an MCP server that exposes framework best-practice patterns as tools, so the coding assistant fetches current patterns on demand URL: https://agentic-patterns.com/patterns/mcp-pattern-injection Markdown: https://agentic-patterns.com/patterns/mcp-pattern-injection.md ### memory-reinforcement-learning-memrl Memory Reinforcement Learning (MemRL): Stores memories with learned utility scores, retrieves by similarity then re-ranks by utility, and updates scores from outcomes while the LLM stays frozen URL: https://agentic-patterns.com/patterns/memory-reinforcement-learning-memrl Markdown: https://agentic-patterns.com/patterns/memory-reinforcement-learning-memrl.md ### memory-synthesis-from-execution-logs Memory Synthesis from Execution Logs: Has the agent write a structured diary per task, then runs synthesis agents over many diaries to turn recurring patterns into rules, commands, and tests URL: https://agentic-patterns.com/patterns/memory-synthesis-from-execution-logs Markdown: https://agentic-patterns.com/patterns/memory-synthesis-from-execution-logs.md ### merged-code-language-skill-model Merged Code + Language Skill Model: Fine-tunes separate language and code specialists from the same base model, then merges their weights into one model instead of one large joint training run URL: https://agentic-patterns.com/patterns/merged-code-language-skill-model Markdown: https://agentic-patterns.com/patterns/merged-code-language-skill-model.md ### milestone-escrow-agent-funding Milestone Escrow for Agent Resource Funding: Holds agent funding in escrow and releases each payment only after independent verification of a measurable milestone URL: https://agentic-patterns.com/patterns/milestone-escrow-agent-funding Markdown: https://agentic-patterns.com/patterns/milestone-escrow-agent-funding.md ### multi-model-orchestration-for-complex-edits Multi-Model Orchestration for Complex Edits: Splits complex code edits across specialized models: a retrieval model gathers context, a large model writes the changes, and smaller models apply them URL: https://agentic-patterns.com/patterns/multi-model-orchestration-for-complex-edits Markdown: https://agentic-patterns.com/patterns/multi-model-orchestration-for-complex-edits.md ### multi-platform-communication-aggregation Multi-Platform Communication Aggregation: Queries every communication platform in parallel through adapters that share one schema, then merges, deduplicates, and ranks the results URL: https://agentic-patterns.com/patterns/multi-platform-communication-aggregation Markdown: https://agentic-patterns.com/patterns/multi-platform-communication-aggregation.md ### multi-platform-webhook-triggers Multi-Platform Webhook Triggers: Starts agent workflows from Slack, Notion, and Jira webhooks, emoji reactions, and schedules, with idempotency and signature checks on each event URL: https://agentic-patterns.com/patterns/multi-platform-webhook-triggers Markdown: https://agentic-patterns.com/patterns/multi-platform-webhook-triggers.md ### no-token-limit-magic No-Token-Limit Magic: Removes hard token limits during prototyping to learn what good behavior needs, then compresses context only after quality is stable and measured URL: https://agentic-patterns.com/patterns/no-token-limit-magic Markdown: https://agentic-patterns.com/patterns/no-token-limit-magic.md ### non-custodial-spending-controls Non-Custodial Spending Controls: Puts a policy layer between the agent and the transaction signer that checks each intent against allowlists, budgets, and rate limits, and fails closed URL: https://agentic-patterns.com/patterns/non-custodial-spending-controls Markdown: https://agentic-patterns.com/patterns/non-custodial-spending-controls.md ### non-generative-judgment-routing Non-Generative Judgment Routing with Typed Escalation: Sends decision-only loop steps as batched typed questions to a non-generative judgment model and escalates what it cannot answer back to the LLM URL: https://agentic-patterns.com/patterns/non-generative-judgment-routing Markdown: https://agentic-patterns.com/patterns/non-generative-judgment-routing.md ### one-os-user-per-agent One OS User per Agent: Give each long-lived agent its own OS user account and a templated init-system unit, so identity, supervision and privilege come from the host instead of from a per-agent container. URL: https://agentic-patterns.com/patterns/one-os-user-per-agent Markdown: https://agentic-patterns.com/patterns/one-os-user-per-agent.md ### opponent-processor-multi-agent-debate Opponent Processor / Multi-Agent Debate Pattern: Spawns agents with opposing roles on the same context, lets them critique each other, then synthesizes their positions to expose bias and blind spots URL: https://agentic-patterns.com/patterns/opponent-processor-multi-agent-debate Markdown: https://agentic-patterns.com/patterns/opponent-processor-multi-agent-debate.md ### oracle-and-worker-multi-model Oracle and Worker Multi-Model Approach: Uses a fast, low-cost worker model for most tool use and code generation, and lets it consult an expensive oracle model when it is stuck URL: https://agentic-patterns.com/patterns/oracle-and-worker-multi-model Markdown: https://agentic-patterns.com/patterns/oracle-and-worker-multi-model.md ### orchestration-prompt-writing-benchmark Orchestration Prompt-Writing Benchmark: Scores an orchestrator on whether its sub-agent prompts give the right information fragments to the right roles, separate from end-task success URL: https://agentic-patterns.com/patterns/orchestration-prompt-writing-benchmark Markdown: https://agentic-patterns.com/patterns/orchestration-prompt-writing-benchmark.md ### out-of-process-provider-boundary-replay Out-of-Process Provider-Boundary Replay: Capture an agent run at the model-provider HTTP boundary from outside the process, then serve those bytes back so the agent re-executes its own logic with no provider call. URL: https://agentic-patterns.com/patterns/out-of-process-provider-boundary-replay Markdown: https://agentic-patterns.com/patterns/out-of-process-provider-boundary-replay.md ### output-verification-loop Output Verification Loop: Verify LLM outputs by extracting individual claims, checking each against evidence sources, and returning per-claim trust scores before acting on the result. URL: https://agentic-patterns.com/patterns/output-verification-loop Markdown: https://agentic-patterns.com/patterns/output-verification-loop.md ### own-check-fault-injection Own-Check Fault Injection: Plant a controlled fault inside a running agent pipeline and score whether the pipeline's own checks emit a detection act, keeping detected, reacted and recovered as separate verdicts. URL: https://agentic-patterns.com/patterns/own-check-fault-injection Markdown: https://agentic-patterns.com/patterns/own-check-fault-injection.md ### parallel-tool-call-learning Parallel Tool Call Learning: Uses agent reinforcement fine-tuning to teach the model to issue independent tool calls in parallel, which cuts sequential rounds and latency URL: https://agentic-patterns.com/patterns/parallel-tool-call-learning Markdown: https://agentic-patterns.com/patterns/parallel-tool-call-learning.md ### patch-steering-via-prompted-tool-selection Patch Steering via Prompted Tool Selection: Tells the agent in the prompt which patch tool or edit format to use, with usage examples, negative rules, and a fallback order URL: https://agentic-patterns.com/patterns/patch-steering-via-prompted-tool-selection Markdown: https://agentic-patterns.com/patterns/patch-steering-via-prompted-tool-selection.md ### persistent-test-memory-feedback-loop Persistent Test Memory Feedback Loop: Preserve validated test lessons so an agent can reuse successful paths, recognize recurring failures, and retire stale knowledge. URL: https://agentic-patterns.com/patterns/persistent-test-memory-feedback-loop Markdown: https://agentic-patterns.com/patterns/persistent-test-memory-feedback-loop.md ### pii-tokenization PII Tokenization: Replaces PII in tool results with placeholder tokens before the model sees them and restores the real values in outgoing tool calls URL: https://agentic-patterns.com/patterns/pii-tokenization Markdown: https://agentic-patterns.com/patterns/pii-tokenization.md ### plan-then-execute-pattern Plan-Then-Execute Pattern: Has the LLM fix the full sequence of tool calls before it reads untrusted data, then runs that sequence so tool outputs change only parameters URL: https://agentic-patterns.com/patterns/plan-then-execute-pattern Markdown: https://agentic-patterns.com/patterns/plan-then-execute-pattern.md ### planner-worker-separation-for-long-running-agents Planner-Worker Separation for Long-Running Agents: Splits agents into planners that create tasks, workers that complete them in isolation, and a judge that decides each cycle whether to continue URL: https://agentic-patterns.com/patterns/planner-worker-separation-for-long-running-agents Markdown: https://agentic-patterns.com/patterns/planner-worker-separation-for-long-running-agents.md ### policy-gated-tool-proxy Policy-Gated Tool Proxy: Insert a transparent proxy between agents and tool servers that evaluates every tool call against a policy engine before forwarding, producing an immutable audit trail of all decisions. URL: https://agentic-patterns.com/patterns/policy-gated-tool-proxy Markdown: https://agentic-patterns.com/patterns/policy-gated-tool-proxy.md ### precomputed-code-graph-lookup Precomputed Code Graph Lookup: Answer an agent's structural code questions from a precomputed, deterministic graph queried by intent-shaped verbs, instead of a retrieval-time search-and-read loop. URL: https://agentic-patterns.com/patterns/precomputed-code-graph-lookup Markdown: https://agentic-patterns.com/patterns/precomputed-code-graph-lookup.md ### proactive-agent-state-externalization Proactive Agent State Externalization: Gives agents note templates, completeness checks, and an external memory fallback so self-written notes keep objectives, decisions, and knowledge gaps URL: https://agentic-patterns.com/patterns/proactive-agent-state-externalization Markdown: https://agentic-patterns.com/patterns/proactive-agent-state-externalization.md ### proactive-trigger-vocabulary Proactive Trigger Vocabulary: Gives each skill an explicit, documented list of trigger phrases and patterns so input routes to skills predictably, with optional proactive activation URL: https://agentic-patterns.com/patterns/proactive-trigger-vocabulary Markdown: https://agentic-patterns.com/patterns/proactive-trigger-vocabulary.md ### progressive-autonomy-with-model-evolution Progressive Autonomy with Model Evolution: Audits prompts and orchestration after each model upgrade and removes the scaffolding that evals show the new model no longer needs URL: https://agentic-patterns.com/patterns/progressive-autonomy-with-model-evolution Markdown: https://agentic-patterns.com/patterns/progressive-autonomy-with-model-evolution.md ### progressive-complexity-escalation Progressive Complexity Escalation: Deploys agents on low-complexity, high-reliability tasks first and unlocks higher capability tiers when metrics and human review gates show proven reliability URL: https://agentic-patterns.com/patterns/progressive-complexity-escalation Markdown: https://agentic-patterns.com/patterns/progressive-complexity-escalation.md ### progressive-disclosure-large-files Progressive Disclosure for Large Files: Puts only file metadata in the prompt and gives the agent load, peek, and extract tools to pull file content into context on demand URL: https://agentic-patterns.com/patterns/progressive-disclosure-large-files Markdown: https://agentic-patterns.com/patterns/progressive-disclosure-large-files.md ### progressive-tool-discovery Progressive Tool Discovery: Organizes tools in a browsable hierarchy and lets the agent load names, descriptions, or full schemas only for the tools it needs URL: https://agentic-patterns.com/patterns/progressive-tool-discovery Markdown: https://agentic-patterns.com/patterns/progressive-tool-discovery.md ### prompt-caching-via-exact-prefix-preservation Prompt Caching via Exact Prefix Preservation: Keeps static prompt content first in a fixed order and only appends new messages, including config changes, so each request reuses the cached prefix URL: https://agentic-patterns.com/patterns/prompt-caching-via-exact-prefix-preservation Markdown: https://agentic-patterns.com/patterns/prompt-caching-via-exact-prefix-preservation.md ### reasoning-token-firewall Reasoning-Token Firewall: Assemble an agent's result only from answer-typed stream events, never by string-stripping interleaved reasoning tokens. URL: https://agentic-patterns.com/patterns/reasoning-token-firewall Markdown: https://agentic-patterns.com/patterns/reasoning-token-firewall.md ### recursive-best-of-n-delegation Recursive Best-of-N Delegation: Runs several parallel candidate workers per subtask in a recursive agent tree, scores them with tests and a judge, and promotes the best result upward URL: https://agentic-patterns.com/patterns/recursive-best-of-n-delegation Markdown: https://agentic-patterns.com/patterns/recursive-best-of-n-delegation.md ### reflection Reflection Loop: Scores each draft against a fixed rubric, feeds the critique into a revision, and repeats until the draft passes a threshold or the retry budget ends URL: https://agentic-patterns.com/patterns/reflection Markdown: https://agentic-patterns.com/patterns/reflection.md ### wfgy-reliability-problem-map Reliability Problem Map Checklist for RAG and Agents: Runs a fixed 16-question failure checklist on each RAG or agent incident, maps the result to repair actions, and re-tests the same failing case URL: https://agentic-patterns.com/patterns/wfgy-reliability-problem-map Markdown: https://agentic-patterns.com/patterns/wfgy-reliability-problem-map.md ### rendered-ui-finish-gate Rendered UI Finish Gate: Verify an agent-built interface against its intended design contract, required states, interaction semantics, and rendered output before merging. URL: https://agentic-patterns.com/patterns/rendered-ui-finish-gate Markdown: https://agentic-patterns.com/patterns/rendered-ui-finish-gate.md ### rich-feedback-loops Rich Feedback Loops > Perfect Prompts: Returns compiler errors, test failures, lint output, and human feedback to the agent after each tool call so it can plan fixes and self-correct URL: https://agentic-patterns.com/patterns/rich-feedback-loops Markdown: https://agentic-patterns.com/patterns/rich-feedback-loops.md ### rlaif-reinforcement-learning-from-ai-feedback RLAIF (Reinforcement Learning from AI Feedback): Uses an AI model guided by written principles to critique outputs and label preferences, which train a reward model that optimizes the policy URL: https://agentic-patterns.com/patterns/rlaif-reinforcement-learning-from-ai-feedback Markdown: https://agentic-patterns.com/patterns/rlaif-reinforcement-learning-from-ai-feedback.md ### sandboxed-tool-authorization Sandboxed Tool Authorization: Filters an agent's tools through deny-first allow and deny patterns, profile presets, and subagent policies that inherit parent restrictions URL: https://agentic-patterns.com/patterns/sandboxed-tool-authorization Markdown: https://agentic-patterns.com/patterns/sandboxed-tool-authorization.md ### schema-validation-retry-cross-step-learning Schema Validation Retry with Cross-Step Learning: Retries failed structured outputs with the validation errors as feedback and adds recent errors from earlier steps to later prompts so mistakes do not repeat URL: https://agentic-patterns.com/patterns/schema-validation-retry-cross-step-learning Markdown: https://agentic-patterns.com/patterns/schema-validation-retry-cross-step-learning.md ### schema-guided-graph-retrieval Schema-Guided Graph Retrieval for Multi-Hop Reasoning: Use one shared domain schema to align graph construction, schema evolution, query decomposition, and typed retrieval so multi-hop reasoning over private knowledge stays precise as domains change. URL: https://agentic-patterns.com/patterns/schema-guided-graph-retrieval Markdown: https://agentic-patterns.com/patterns/schema-guided-graph-retrieval.md ### seamless-background-to-foreground-handoff Seamless Background-to-Foreground Handoff: Lets a user take over a background agent's unfinished work in the foreground, with the agent's branch, PR, and summaries carried over as context URL: https://agentic-patterns.com/patterns/seamless-background-to-foreground-handoff Markdown: https://agentic-patterns.com/patterns/seamless-background-to-foreground-handoff.md ### self-critique-evaluator-loop Self-Critique Evaluator Loop: Trains a judge model on its own synthetic comparisons of candidate outputs and uses it as a reward model or quality gate for the main agent URL: https://agentic-patterns.com/patterns/self-critique-evaluator-loop Markdown: https://agentic-patterns.com/patterns/self-critique-evaluator-loop.md ### self-discover-reasoning-structures Self-Discover: LLM Self-Composed Reasoning Structures: Has the LLM select, adapt, and compose reasoning modules into a task-specific reasoning structure, then solve the task by following that structure URL: https://agentic-patterns.com/patterns/self-discover-reasoning-structures Markdown: https://agentic-patterns.com/patterns/self-discover-reasoning-structures.md ### self-identity-accumulation Self-Identity Accumulation: Injects a persistent identity document at session start and updates it with new user insights at session end through lifecycle hooks URL: https://agentic-patterns.com/patterns/self-identity-accumulation Markdown: https://agentic-patterns.com/patterns/self-identity-accumulation.md ### self-rewriting-meta-prompt-loop Self-Rewriting Meta-Prompt Loop: Has the agent reflect after each episode, draft edits to its own system prompt, validate them through guardrails, and save the new version URL: https://agentic-patterns.com/patterns/self-rewriting-meta-prompt-loop Markdown: https://agentic-patterns.com/patterns/self-rewriting-meta-prompt-loop.md ### semantic-context-filtering Semantic Context Filtering Pattern: Extracts only the semantic or interactive parts of raw data, such as accessibility trees or relevant fields, before sending it to the LLM URL: https://agentic-patterns.com/patterns/semantic-context-filtering Markdown: https://agentic-patterns.com/patterns/semantic-context-filtering.md ### session-scoped-context-runtime-for-agent-tools Session-Scoped Context Runtime for Agent Tools: Interpose a context runtime that caches structured reads and normalizes tool output so sessions reuse compact representations instead of repeating raw tokens. URL: https://agentic-patterns.com/patterns/session-scoped-context-runtime-for-agent-tools Markdown: https://agentic-patterns.com/patterns/session-scoped-context-runtime-for-agent-tools.md ### shell-command-contextualization Shell Command Contextualization: Lets the user run a shell command with a prefix such as ! and injects the command and its full output into the agent's context URL: https://agentic-patterns.com/patterns/shell-command-contextualization Markdown: https://agentic-patterns.com/patterns/shell-command-contextualization.md ### shipping-as-research Shipping as Research: Releases reversible, instrumented features to learn whether they work, then doubles down or removes them based on usage data and feedback URL: https://agentic-patterns.com/patterns/shipping-as-research Markdown: https://agentic-patterns.com/patterns/shipping-as-research.md ### signal-driven-agent-activation Signal-Driven Agent Activation: Watches external sources for structured signals and starts predefined agent workflows when declarative rules with thresholds and cooldowns match URL: https://agentic-patterns.com/patterns/signal-driven-agent-activation Markdown: https://agentic-patterns.com/patterns/signal-driven-agent-activation.md ### skill-activation-as-precision-recall Skill Activation as a Precision/Recall Measurement: Measures skill routing as classification with a labelled prompt dataset, per-skill precision and recall, and CI thresholds that fail the build URL: https://agentic-patterns.com/patterns/skill-activation-as-precision-recall Markdown: https://agentic-patterns.com/patterns/skill-activation-as-precision-recall.md ### skill-library-evolution Skill Library Evolution: Saves working agent code as reusable skills in a skills directory, documents and tests them over time, and loads skill details only on demand URL: https://agentic-patterns.com/patterns/skill-library-evolution Markdown: https://agentic-patterns.com/patterns/skill-library-evolution.md ### soulbound-identity-verification Soulbound Identity Verification: Binds agent identity to a non-transferable credential with a committed state hash and logs signed state changes so verifiers can check continuity URL: https://agentic-patterns.com/patterns/soulbound-identity-verification Markdown: https://agentic-patterns.com/patterns/soulbound-identity-verification.md ### spec-as-test-feedback-loop Spec-As-Test Feedback Loop: Generates executable tests from the spec on every spec or code commit and opens agent PRs that fix code or flag unclear spec parts URL: https://agentic-patterns.com/patterns/spec-as-test-feedback-loop Markdown: https://agentic-patterns.com/patterns/spec-as-test-feedback-loop.md ### specification-driven-agent-development Specification-Driven Agent Development: Makes a version-controlled spec file the agent's main input, scaffolds code from it, and links each artifact back to a spec clause URL: https://agentic-patterns.com/patterns/specification-driven-agent-development Markdown: https://agentic-patterns.com/patterns/specification-driven-agent-development.md ### spectrum-of-control-blended-initiative Spectrum of Control / Blended Initiative: Gives users several autonomy modes, from inline completion to background agents, and lets them switch modes per task URL: https://agentic-patterns.com/patterns/spectrum-of-control-blended-initiative Markdown: https://agentic-patterns.com/patterns/spectrum-of-control-blended-initiative.md ### static-service-manifest-for-agents Static Service Manifest for Agents: Serves a static llms.txt or agent.json file at a well-known URL that lists services, auth, and limits so agents can plan before they call URL: https://agentic-patterns.com/patterns/static-service-manifest-for-agents Markdown: https://agentic-patterns.com/patterns/static-service-manifest-for-agents.md ### stop-hook-auto-continue-pattern Stop Hook Auto-Continue Pattern: Runs a stop hook after each agent turn that checks success criteria and makes the agent continue until tests or checks pass URL: https://agentic-patterns.com/patterns/stop-hook-auto-continue-pattern Markdown: https://agentic-patterns.com/patterns/stop-hook-auto-continue-pattern.md ### structured-output-specification Structured Output Specification: Constrain agent outputs using deterministic schemas that enforce structured, machine-readable results, enabling reliable validation, parsing, and integration with downstream systems. URL: https://agentic-patterns.com/patterns/structured-output-specification Markdown: https://agentic-patterns.com/patterns/structured-output-specification.md ### sub-agent-spawning Sub-Agent Spawning: Lets the main agent spawn sub-agents with fresh context and scoped tools to work on subtasks in parallel, then merges their results URL: https://agentic-patterns.com/patterns/sub-agent-spawning Markdown: https://agentic-patterns.com/patterns/sub-agent-spawning.md ### subagent-compilation-checker Subagent Compilation Checker: Spawns one subagent per module to build and check it, and returns only a short structured error list or artifact reference to the main agent URL: https://agentic-patterns.com/patterns/subagent-compilation-checker Markdown: https://agentic-patterns.com/patterns/subagent-compilation-checker.md ### subject-hygiene Subject Hygiene for Task Delegation: Requires a specific action-plus-target subject on every subagent task so each delegated job stays traceable and easy to reference URL: https://agentic-patterns.com/patterns/subject-hygiene Markdown: https://agentic-patterns.com/patterns/subject-hygiene.md ### surface-scoped-system-prompt Surface-Scoped System Prompt: Ships one system prompt per invocation surface and selects it at launch, so an interactive session, a headless run and a task mode each get instructions written for that surface URL: https://agentic-patterns.com/patterns/surface-scoped-system-prompt Markdown: https://agentic-patterns.com/patterns/surface-scoped-system-prompt.md ### swarm-migration-pattern Swarm Migration Pattern: Main agent orchestrates 10+ parallel subagents working simultaneously on independent migration chunks for large-scale framework upgrades, lint rule rollouts, and API migrations. URL: https://agentic-patterns.com/patterns/swarm-migration-pattern Markdown: https://agentic-patterns.com/patterns/swarm-migration-pattern.md ### team-shared-agent-configuration Team-Shared Agent Configuration as Code: Check agent configuration into version control as code, enabling consistent behavior across teams, faster onboarding, and collaborative improvement through PRs and code review. URL: https://agentic-patterns.com/patterns/team-shared-agent-configuration Markdown: https://agentic-patterns.com/patterns/team-shared-agent-configuration.md ### three-stage-perception-architecture Three-Stage Perception Architecture: Splits the agent into separate perception, processing, and action stages so each stage can be built, tested, and scaled on its own URL: https://agentic-patterns.com/patterns/three-stage-perception-architecture Markdown: https://agentic-patterns.com/patterns/three-stage-perception-architecture.md ### tier-auto-apply-by-mechanical-impact Tier Auto-Apply by Mechanical Impact: Decides which self-proposed changes auto-apply by matching the real diff's file paths and change kinds against a trusted tier table, not by finding text URL: https://agentic-patterns.com/patterns/tier-auto-apply-by-mechanical-impact Markdown: https://agentic-patterns.com/patterns/tier-auto-apply-by-mechanical-impact.md ### tool-capability-compartmentalization Tool Capability Compartmentalization: Splits tools into reader, processor, and writer classes with scoped permissions and blocks tool chains that combine private data, untrusted input, and external writes URL: https://agentic-patterns.com/patterns/tool-capability-compartmentalization Markdown: https://agentic-patterns.com/patterns/tool-capability-compartmentalization.md ### tool-search-lazy-loading Tool Search Lazy Loading: Dynamically load tools via search instead of preloading all available tools to reduce context usage URL: https://agentic-patterns.com/patterns/tool-search-lazy-loading Markdown: https://agentic-patterns.com/patterns/tool-search-lazy-loading.md ### tool-selection-guide Tool Selection Guide: Maps each task type to a preferred tool: Glob, Grep, and Read to explore, Edit to modify, Bash to verify, and Task with a clear subject to delegate URL: https://agentic-patterns.com/patterns/tool-selection-guide Markdown: https://agentic-patterns.com/patterns/tool-selection-guide.md ### tool-use-incentivization-via-reward-shaping Tool Use Incentivization via Reward Shaping: Gives dense RL rewards for useful intermediate tool calls such as compile, lint, and test so the agent learns to use tools instead of only thinking URL: https://agentic-patterns.com/patterns/tool-use-incentivization-via-reward-shaping Markdown: https://agentic-patterns.com/patterns/tool-use-incentivization-via-reward-shaping.md ### tool-use-steering-via-prompting Tool Use Steering via Prompting: Tells the agent in the prompt which tool to use, how to learn a custom tool, and which shorthands map to tool sequences URL: https://agentic-patterns.com/patterns/tool-use-steering-via-prompting Markdown: https://agentic-patterns.com/patterns/tool-use-steering-via-prompting.md ### tracker-as-desired-state-reconciliation Tracker-as-Desired-State Reconciliation: Derive agent dispatch by re-reading workflow states in the team's issue tracker every tick, keeping only a short-lived claim locally, so missed events, restarts, and human edits all self-heal. URL: https://agentic-patterns.com/patterns/tracker-as-desired-state-reconciliation Markdown: https://agentic-patterns.com/patterns/tracker-as-desired-state-reconciliation.md ### transitive-vouch-chain-trust Transitive Vouch-Chain Trust: Builds a graph of signed vouches between agents and derives trust in an unknown agent from the chain back to a trusted one, with decay at each hop URL: https://agentic-patterns.com/patterns/transitive-vouch-chain-trust Markdown: https://agentic-patterns.com/patterns/transitive-vouch-chain-trust.md ### tree-of-thought-reasoning Tree-of-Thought Reasoning: Expands a tree of candidate reasoning steps, scores partial states, prunes weak branches, and picks the best path instead of one linear chain URL: https://agentic-patterns.com/patterns/tree-of-thought-reasoning Markdown: https://agentic-patterns.com/patterns/tree-of-thought-reasoning.md ### unified-tool-gateway Unified Tool Gateway: Route all agent tool calls through a single gateway that handles discovery, authentication, billing, and execution across many heterogeneous providers. URL: https://agentic-patterns.com/patterns/unified-tool-gateway Markdown: https://agentic-patterns.com/patterns/unified-tool-gateway.md ### variance-based-rl-sample-selection Variance-Based RL Sample Selection: Runs the base model several times per sample and trains RL only on samples with score variance, skipping ones that are always right or always wrong URL: https://agentic-patterns.com/patterns/variance-based-rl-sample-selection Markdown: https://agentic-patterns.com/patterns/variance-based-rl-sample-selection.md ### verbose-reasoning-transparency Verbose Reasoning Transparency: Lets users open a verbose view on demand that shows the agent's interpretation, tool choices, intermediate steps, and raw tool outputs URL: https://agentic-patterns.com/patterns/verbose-reasoning-transparency Markdown: https://agentic-patterns.com/patterns/verbose-reasoning-transparency.md ### versioned-constitution-governance Versioned Constitution Governance: Stores the agent constitution in a signed Git repository where the agent can only propose changes and reviewers or CI gates merge them URL: https://agentic-patterns.com/patterns/versioned-constitution-governance Markdown: https://agentic-patterns.com/patterns/versioned-constitution-governance.md ### virtual-machine-operator-agent Virtual Machine Operator Agent: Gives the agent a dedicated virtual machine where it can run code, install packages, use the file system, and operate CLI tools URL: https://agentic-patterns.com/patterns/virtual-machine-operator-agent Markdown: https://agentic-patterns.com/patterns/virtual-machine-operator-agent.md ### visual-ai-multimodal-integration Visual AI Multimodal Integration: Adds multimodal models to the agent so it can analyze images, video, and screenshots and combine them with text to reason and act URL: https://agentic-patterns.com/patterns/visual-ai-multimodal-integration Markdown: https://agentic-patterns.com/patterns/visual-ai-multimodal-integration.md ### workflow-evals-with-mocked-tools Workflow Evals with Mocked Tools: Runs complete agent workflows against mocked tools in CI and checks which tools were called plus agent-as-judge quality criteria URL: https://agentic-patterns.com/patterns/workflow-evals-with-mocked-tools Markdown: https://agentic-patterns.com/patterns/workflow-evals-with-mocked-tools.md ### working-memory-via-todos Working Memory via Task Tracking: Keeps an explicit task list with status, blockers, and next steps during the session so agent and user can track progress URL: https://agentic-patterns.com/patterns/working-memory-via-todos Markdown: https://agentic-patterns.com/patterns/working-memory-via-todos.md ### workspace-native-multi-agent-orchestration Workspace-Native Multi-Agent Orchestration: Runs agents inside the team's workspace platform so they share its memory, knowledge sources, event triggers, and integrations with humans URL: https://agentic-patterns.com/patterns/workspace-native-multi-agent-orchestration Markdown: https://agentic-patterns.com/patterns/workspace-native-multi-agent-orchestration.md ### zero-knowledge-verified-agent-egress Zero-Knowledge Verified Agent Egress: Intercepts each outbound HTTP or MCP call, proves its claims against a private source of truth with a zero-knowledge proof, and blocks calls that fail URL: https://agentic-patterns.com/patterns/zero-knowledge-verified-agent-egress Markdown: https://agentic-patterns.com/patterns/zero-knowledge-verified-agent-egress.md ### zero-trust-agent-mesh Zero-Trust Agent Mesh: Gives each agent a cryptographic identity and verifies identity, signed delegation tokens, and chain depth on every inter-agent request URL: https://agentic-patterns.com/patterns/zero-trust-agent-mesh Markdown: https://agentic-patterns.com/patterns/zero-trust-agent-mesh.md