Prompt Engineering Projects - Expanded Guides
Source guide:
../PROMPT_ENGINEERING_PROJECTS.md
This folder contains expanded, implementation-ready files for each project in the Prompt Engineering mastery sprint.
Project Index
| # | Project | File | Outcome |
|---|---|---|---|
| 1 | Prompt Contract Harness | P01-prompt-contract-harness.md | A contract-test harness that validates prompt responses against typed invariants and release thresholds. |
| 2 | JSON Output Enforcer (Schema + Repair Loop) | P02-json-output-enforcer.md | A structured-output gateway that enforces JSON schemas and runs bounded repair retries. |
| 3 | Prompt Injection Red-Team Lab | P03-prompt-injection-red-team-lab.md | A red-team simulator that runs direct/indirect injection attacks and scores policy containment. |
| 4 | Context Window Manager | P04-context-window-manager.md | A context-packing pipeline that retrieves, reranks, compresses, and packs evidence within token budgets. |
| 5 | Few-Shot Example Curator | P05-few-shot-example-curator.md | An example-bank curation and selection pipeline that maximizes quality lift while minimizing bias drift. |
| 6 | Tool Router (Function Schemas as Contracts) | P06-tool-router.md | A runtime tool-router API that maps intent to tool calls with strict schema and policy gates. |
| 7 | Temperature Sweeper + Confidence Policy | P07-temperature-sweeper-confidence-policy.md | A reliability experiment runner that sweeps sampling settings and learns safe confidence policies. |
| 8 | Prompt DSL + Linter | P08-prompt-dsl-linter.md | A domain-specific prompt language and linter that enforces style, safety, and maintainability rules in CI. |
| 9 | Prompt Caching Optimizer | P09-prompt-caching-optimizer.md | A benchmark harness that redesigns prompt prefixes for cache-hit improvements and cost reduction. |
| 10 | Citation Grounding Gateway | P10-citation-grounding-gateway.md | An API gateway that blocks ungrounded claims and returns only citation-backed responses. |
| 11 | Canary Prompt Rollout Controller | P11-canary-prompt-rollout-controller.md | A rollout controller that shifts traffic between prompt versions and auto-promotes or rolls back by policy. |
| 12 | Conversation Memory Compressor | P12-conversation-memory-compressor.md | A memory subsystem that compresses long conversations into policy-safe summaries with recall guarantees. |
| 13 | Tool Permission Firewall | P13-tool-permission-firewall.md | A policy firewall that enforces capability-based permissions and human approval for risky tool actions. |
| 14 | Adversarial Eval Forge | P14-adversarial-eval-forge.md | A continuous adversarial evaluation forge that mutates attacks and tracks security trends over time. |
| 15 | Prompt Registry + Versioning Service | P15-prompt-registry-versioning-service.md | A registry API for versioned prompt artifacts, compatibility checks, approvals, and audit history. |
| 16 | Human-in-the-Loop Escalation Queue | P16-human-in-the-loop-escalation-queue.md | A reviewer workflow system that escalates uncertain/high-risk model outputs and tracks SLA + override quality. |
| 17 | MCP Contract Verifier | P17-mcp-contract-verifier.md | A conformance test suite that validates MCP tool/resource behavior against declared contracts and policies. |
| 18 | Production Prompt Platform Capstone | P18-production-prompt-platform-capstone.md | An end-to-end PromptOps control plane that integrates contracts, evals, rollouts, routing, and incident response. |
| 19 | Reasoning Modes and Self-Verification Lab | P19-reasoning-modes-self-verification-lab.md | A policy harness that controls reasoning depth, self-verification passes, and confidence-based escalation. |
| 20 | Context Construction and Memory Control Plane | P20-context-construction-memory-control-plane.md | A context assembly system for retrieval packing, memory pruning, and deterministic token budgeting. |
| 21 | Tool Selection and Reliability Harness | P21-tool-selection-reliability-harness.md | A tool-calling reliability lab with arbitration prompts, validation gates, and verifier-based retries. |
| 22 | Multi-Agent Orchestration Studio | P22-multi-agent-orchestration-studio.md | An architecture simulator comparing single-agent, supervisor-worker, and parallel-voting orchestration patterns. |
| 23 | Prompt Benchmark and Scoring Factory | P23-prompt-benchmark-scoring-factory.md | A benchmark pipeline for golden sets, rubric scoring, drift detection, and release regression gates. |
| 24 | Injection and Jailbreak Defense Gauntlet | P24-injection-jailbreak-defense-gauntlet.md | A layered adversarial defense suite for prompt injection, jailbreak, and context poisoning scenarios. |
| 25 | Provider-Specific Prompt Optimizer | P25-provider-specific-prompt-optimizer.md | A canonical prompt compiler that emits provider-tuned variants and compares quality/cost/latency tradeoffs. |
| 26 | Cost-Latency Routing Controller | P26-cost-latency-routing-controller.md | A multi-model routing controller for prompt minimization and cheap-to-expensive escalation policies. |
| 27 | Multimodal Agent Prompting Studio | P27-multimodal-agent-prompting-studio.md | A vision-plus-tool workflow lab with structured uncertainty handling for image-grounded tasks. |
| 28 | Cognitive Prompt Architecture Workbench | P28-cognitive-prompt-architecture-workbench.md | A stateful prompting framework for working memory, goal stacks, and metacognitive checks. |
| 29 | Agent Failure Recovery Orchestrator | P29-agent-failure-recovery-orchestrator.md | A recovery engine with retry-strategy mutation, fallback model routing, and state-reset controls. |
| 30 | Prompt DSL and Composition Engine | P30-prompt-dsl-composition-engine.md | A declarative prompt DSL with templates, macros, composition validation, and semantic versioning. |
| 31 | Safety-Critical Prompt Gatekeeper | P31-safety-critical-prompt-gatekeeper.md | A safety policy gate for risk classification, irreversibility detection, and mandatory human approval flows. |
| 32 | Real-World Agent Architecture Casebook | P32-real-world-agent-architecture-casebook.md | A four-architecture portfolio covering SaaS support, research RAG, code PEC, and autonomous ops agents. |
| 33 | Prompt Anti-Pattern Detection Lab | P33-prompt-anti-pattern-detection-lab.md | A linter and drift monitor for over-constraining, verbosity, context flooding, redundancy, and prompt drift. |
Notes
- Each file now includes concrete project specifications, deterministic real-world outcomes, and implementation/testing guidance.
- Use these files with
../PROMPT_ENGINEERING_PROJECTS.mdto follow the theory-first + build-second workflow.