Prompt Library
204 Prompts — Alle Kategorien
Matcher — Top-3 aehnliche Prompts
Improver — Prompt verbessern
| Name | Kategorie | Vorschau | Variablen | Rating | Modell | Tags |
|---|---|---|---|---|---|---|
| FastAPI Async Endpoint | Code | You are a senior Python backend engineer. Design a production-ready FastAPI endpoint for: {{feature}} Requirements: - Use async/await throughout; no | feature, context | ★★★★★ | ollama/qwen2.5-coder:7b | python,fastapi,async,pydantic,sqlalchemy |
| Investigate glug_lessons Contradiction Signal | Analysis | Trace and resolve a contradiction signal for lesson {lesson_id} in the Knowledge Governance Board. Access point: https://orchestrator.joelduss.xyz/go | lesson_id, search_query, topic, winner_id | ★★★★★ | openai/gpt-4.1 | stack-aware,rag-stack,lessons,contradiction,ap11,governance,knowledge |
| Extend Glug Addon Distribution API | Code | Add support for a new addon type `{addon_type}` to the glug-server addon-distribution API. Existing API: /api/v1/glug/addons/* (registered in glug we | addon_type | ★★★★★ | openai/gpt-4.1 | stack-aware,glug,addon,api,strict-separation,plugin-registry |
| Diagnose Flight Recorder AP4 Anomaly | Analysis | Diagnose an anomalous entry in the AP4 tamper-evident flight recorder audit log. Log location: /var/log/agent-flight-recorder.jsonl (append-only JSON | anomaly_entry, entry_id | ★★★★★ | openai/gpt-4.1 | stack-aware,rag-stack,flight-recorder,ap4,audit,tamper-detection,security |
| Audit Yoga Glug-Uplink Imports | Analysis | Audit the yoga codebase at {repo_path} for strict HTTP-only separation from glug-server. Separation rule: yoga ↔ glug communication MUST be HTTP-only | repo_path | ★★★★★ | openai/gpt-4.1 | stack-aware,yoga,glug,import,strict-separation,audit,security |
| Yoga Tenant Override Setup | System | Set up a tenant-specific override in yoga for tenant {tenant_id}. Yoga's tenant-isolation rules (from AGENTS.md §14): - Code IS shared across contain | tenant_id, template_name, slug, page_title | ★★★★★ | openai/gpt-4.1 | stack-aware,yoga,tenant,isolation,override,plugin-bootstrap |
| Plan Wave-Style Migration | Flow | Design a Wave-style migration to move {feature_name} from {source_service} to {target_service}. Pattern (from rag-stack Wave-1 through Wave-8d): - Ea | feature_name, source_service, target_service, wave_number | ★★★★★ | openai/gpt-4.1 | stack-aware,migration,wave,glug,yoga,rag-stack,strict-separation |
| Author Glug or Yoga ADR | Text | Draft a new Architecture Decision Record (ADR) for a decision affecting glug-server or yoga. File location: docs/adr/{adr_number}-{slug}.md (e.g. 004 | adr_number, slug, status, change_summary, decision_text, positive_consequences, negative_consequences | ★★★★★ | openai/gpt-4.1 | stack-aware,adr,docs,glug,yoga,architecture,decision |
| Obsidian RAG Watch n8n Integration | Flow | Connect the obsidian-rag-watch service to n8n for automated lesson ingestion. Architecture: - n8n-rag-watch.service (systemd unit) watches /nvmetank1 | vault_folder_pattern | ★★★★★ | openai/gpt-4.1 | stack-aware,n8n,obsidian,rag-watch,lessons,brain,systemd |
| n8n Workflow with Project Tag | Flow | Design an n8n workflow tagged `project:{repo}` that automates {automation_task} for the {repo} Gitea repo. Workflow structure: 1. Trigger: Webhook no | repo, automation_task, branch, extra_tags | ★★★★★ | openai/gpt-4.1 | stack-aware,n8n,workflow,project-tag,gitea,rag-stack |
| Agent Token Rotation Procedure | System | Rotate the Gitea token for agent role {agent_role} stored at /etc/gitea-tokens/{agent_role}.token. Procedure (per AP3 secrets policy): 1. Generate ne | agent_role, date | ★★★★★ | openai/gpt-4.1 | stack-aware,gitea,token,rotation,security,ap3 |
| Caddy Vhost + Pi-hole + Authelia | System | Wire a new service {service_name} running on port {port} through the full stack: Caddy + Pi-hole + Authelia. Step 1 — Pi-hole v6 DNS: Edit /etc/pihol | service_name, port | ★★★★★ | openai/gpt-4.1 | stack-aware,caddy,pi-hole,authelia,reverse-proxy,dns |
| Agent Quality Supervisor Drift Signal | Analysis | Diagnose a drift signal detected by bin/agent-quality-supervisor for agent role {agent_role}. Context: - agent-quality-supervisor implements AP50.5 1 | agent_role, drift_categories | ★★★★★ | openai/gpt-4.1 | stack-aware,rag-stack,agent-quality-supervisor,ap50,drift,governance |
| Agent-Orchestrate Wave Dispatch | Flow | Plan and execute a {wave_count}-wave multi-agent task dispatch using bin/agent-orchestrate. Context: - agent-orchestrate implements AP50 multi-projec | wave_count, task_description, model_a, subtask_a, artifact_a, model_b, subtask_b, artifact_b, model_c | ★★★★★ | openai/gpt-4.1 | stack-aware,rag-stack,agent-orchestrate,ap50,multi-agent,waves |
| RAG Governance Knowledge Page Section | Text | Add a new section to the rag-stack /governance/knowledge/ page following AP56 patterns. The Knowledge Governance Board UI lives in: - Blueprint: src/ | section_name, section_slug, tag_filter | ★★★★★ | openai/gpt-4.1 | stack-aware,rag-stack,governance,ap56,knowledge,lessons |
| AP11 Contradiction Detection Query | Analysis | Investigate a contradiction signal in glug_lessons using the AP11 contradiction-detector. Context: - rag.db at /persistent/rag-stack/rag.db - Tables: | topic | ★★★★★ | openai/gpt-4.1 | stack-aware,rag-stack,ap11,contradiction-detector,lessons,governance |
| RAG Add Lesson with AP3 Secret Scan | Flow | Add a new lesson to rag-stack using the rag-add CLI with AP3 secret-scan compliance. Step-by-step: 1. Verify no secrets in lesson body (AP3 + FU2): a | lesson_body, topic, tags | ★★★★★ | openai/gpt-4.1 | stack-aware,rag-stack,rag-add,ap3,secret-scan,lessons |
| Yoga Plugin PLUGIN Dict Registration | Code | Create a new yoga plugin using the PLUGIN dict pattern (not TOML). Yoga plugins use Python dicts, not manifest.toml. Plugin ID format: yoga.{plugin_i | plugin_id, plugin_display_name, trigger_signal | ★★★★★ | openai/gpt-4.1 | stack-aware,yoga,plugin,eda,tenant-isolation |
| Yoga Blinker Signal Subscriber | Code | Add a new Blinker signal subscriber to yoga's EDA layer at /nvmetank1/docker/yoga/signals.py. Context: yoga has 45 signals and 41 active connections. | signal_name, module, handler_action, expected_connection_count | ★★★★★ | openai/gpt-4.1 | stack-aware,yoga,blinker,signals,eda |
| Glug Plugin EDA Signal Subscription | Code | Wire a Glug plugin at /nvmetank1/projects/glug/plugins/{plugin_name}/ to an EDA signal. Context: Glug uses Blinker signals (signals.py, 45 signals, 4 | plugin_name, signal_name, action | ★★★★★ | openai/gpt-4.1 | stack-aware,glug,plugin,eda,blinker,signals |
| Glug Plugin DB Migration | Code | Add a database migration to a Glug plugin at /nvmetank1/projects/glug/plugins/{plugin_name}/. Requirements: - Add a "db_migrations" key to manifest.t | plugin_name, table_name, extra_columns | ★★★★★ | openai/gpt-4.1 | stack-aware,glug,plugin,db-migration,sqlite |
| Add Glug Plugin with Blueprint | Code | Create a new Glug plugin at /nvmetank1/projects/glug/plugins/{plugin_name}/. Your plugin MUST include: 1. manifest.toml with id = "glug.{plugin_name} | plugin_name, feature | ★★★★★ | openai/gpt-4.1 | stack-aware,glug,plugin,flask,blueprint,manifest |
| Wave-8c-style data migration between SQLite DBs | DevOps | Migrate data from `{source_db}` to `{dest_db}` following Wave-8c (rag-stack#50) idempotent + verifiable pattern. Use `bin/migrate-lessons-from-glug.p | {'name': 'source_db', 'description': 'path'}, {'name': 'dest_db', 'description': 'path'}, {'name': 'topic', 'description': 'slug for script names'}, {'name': 'table_name', 'description': 'per-table'}, {'name': 'tables', 'description': 'list'}, {'name': 'reason', 'description': 'why migrate'} | ★★★★★ | claude-3.7-sonnet | stack-aware,sqlite,migration,wave-8c,backup |
| Route prompt via /api/v1/prompts/match | Code | Use the rag-stack prompt-library to find a high-quality template for `{user_query}` BEFORE writing one from scratch. 1. Search: ```bash curl -sX POST | {'name': 'user_query', 'description': 'raw query in user words'} | ★★★★★ | claude-3.5-haiku | stack-aware,prompts,prompt-library,rag-stack-83 |
| yoga plugin with EDA + tenant-overrides | Code | Add a yoga.{plugin_name} plugin in `/nvmetank1/docker/yoga/plugins/{plugin_name}/` (post-Wave-1+2 separation: yoga lives in its own repo). Files: 1. | {'name': 'plugin_name', 'description': 'snake_case'}, {'name': 'feature', 'description': 'what + why'} | ★★★★★ | claude-3.7-sonnet | stack-aware,yoga,plugin,multi-tenant,eda,strict-separation |
| agent-orchestrate dispatch with AP3 stages | Flow | Plan a multi-project agent dispatch via `bin/agent-orchestrate` for `{intent}` across `{projects}`. Pre-flight (AP3 6-stage policy): 1. Stage 1 (role | {'name': 'intent', 'description': 'long-form intent'}, {'name': 'intent_short', 'description': '≤80 chars'}, {'name': 'projects', 'description': 'description of which projects'}, {'name': 'projects_comma', 'description': 'glug,rag-stack,yoga,…'}, {'name': 'role', 'description': 'agent-haiku|claude|or|codex|qwen'}, {'name': 'parallel_count', 'description': 'max parallel children'}, {'name': 'timeout_min', 'description': 'per-child timeout'} | ★★★★★ | claude-3.7-sonnet | stack-aware,orchestration,agent-orchestrate,ap3,ap4,ap50 |
| rag-add lesson with AP5 source-authority | Code | Author a rag-add lesson for `{topic}` that's well-formed for AP5 source-authority + FU2 secret-scan + obsidian-mirror. CLI invocation (note: rag-add | {'name': 'topic', 'description': 'long topic'}, {'name': 'topic_short', 'description': '≤80 chars'}, {'name': 'tags_comma', 'description': 'comma-separated'}, {'name': 'agent_role', 'description': 'agent-claude/human/etc'}, {'name': 'issue_num', 'description': 'gitea issue#'}, {'name': 'author_kind', 'description': 'tool|human|ai'}, {'name': 'lesson_text', 'description': 'actual lesson body'} | ★★★★★ | claude-3.5-haiku | stack-aware,rag,lessons,ap5,fu2,obsidian |
| FU5 docker.sock audit pattern | Analysis | Audit `{compose_file}` (or specific service `{service}`) for docker.sock exposure following FU5 methodology (rag-stack#84). Steps: 1. `grep -nE 'dock | {'name': 'compose_file', 'description': 'e.g. /docker-yaml/ai-stack.yml'}, {'name': 'service', 'description': 'specific container, optional'} | ★★★★★ | claude-3.7-sonnet | stack-aware,security,docker-sock,fu5,audit |
| Wave-style migration plan glug ↔ rag-stack ↔ yoga | Analysis | Plan a strict-separation migration of `{feature}` between glug/rag-stack/yoga repos. Per user 2026-05-01: glug = web framework only; rag-stack = agen | {'name': 'feature', 'description': 'e.g. agent-quality-supervisor'}, {'name': 'feature_description', 'description': 'what + why migrating'}, {'name': 'source_repo', 'description': 'glug|rag_stack|yoga'}, {'name': 'target_repo', 'description': 'glug|rag_stack|yoga'}, {'name': 'topic', 'description': 'backup-tag slug'} | ★★★★★ | claude-3.7-sonnet | stack-aware,migration,wave,strict-separation,rag-stack-45 |
| Author rag-stack /governance/{page} blueprint | Code | Add a new governance UI page at `/governance/{page}` in rag-stack, modeled after AP56 Knowledge Governance Board. Files to create in `/nvmetank1/proj | {'name': 'page', 'description': 'snake_case slug'}, {'name': 'feature_description', 'description': 'what the board surfaces'} | ★★★★★ | claude-3.7-sonnet | stack-aware,rag-stack,governance,blueprint,ap56 |