Prompt Assembly
March 6, 2026 · View on GitHub
🏥 Quick Return to Emergency Room
You are in a specialist desk.
For full triage and doctors on duty, return here:
- WFGY Global Fix Map — main Emergency Room, 300+ structured fixes
- WFGY Problem Map 1.0 — 16 reproducible failure modes
Think of this page as a sub-room.
If you want full consultation and prescriptions, go back to the Emergency Room lobby.
Build prompts that models cannot misread.
Use this folder when citations vanish, JSON mode breaks, tools loop, or answers flip after a small template change.
Every page gives a concrete repair with measurable targets. No infra change required.
Orientation: what each page does
| Page | What it solves | Typical symptom |
|---|---|---|
| System vs User Role Order | Locks role hierarchy and section order | Role text bleeds into user content, answers flip after reorder |
| JSON Mode and Tool Calls | Validates schemas and fences tool outputs | Free text in tool responses, invalid JSON, missing fields |
| Citation First | Enforces cite then explain with required fields | Citations missing or point to the wrong snippet |
| Anti Prompt Injection Recipes | Ready to paste defenses for common exploits | Hidden instructions override system prompt |
| Memory Fences and State Keys | Prevents cross turn or cross agent overwrite | Agents rewrite each other’s memory, history leaks |
| Tool Selection and Timeouts | Picks tool deterministically, adds timeouts | Loops, stalls, or wrong tool chosen |
| Template Library (minimal) | Small set of reusable prompt blocks | Inconsistent phrasing across agents or runs |
| Eval Prompts and Checks | Deterministic acceptance gates | “Looks better” but no stable way to prove it |
When to use
- Citations point to the wrong snippet or disappear after retries.
- JSON mode produces invalid objects or tool calls stall in loops.
- Role text bleeds into user content after a small template change.
- Long chains smear topics when you reorder headers.
- Agents overwrite each other’s memory without fences.
Acceptance targets
- ΔS(question, retrieved) ≤ 0.45
- Coverage of target section ≥ 0.70
- λ_observe convergent across three paraphrases and two seeds
- E_resonance flat on long windows
Map symptoms to structural fixes
| Symptom | Likely cause | Open this |
|---|---|---|
| Wrong meaning despite high similarity | Metric or analyzer mismatch | Embedding ≠ Semantic |
| Citations inconsistent or missing after retries | No traceability schema enforced | Retrieval Traceability · Data Contracts |
| JSON breaks or tool responses contain free text | Tool schema not fenced, logic collapsed | JSON Mode and Tool Calls · Prompt Injection |
| Answers flip when you reorder headers | Header order changes λ state | System vs User Role Order · Context Drift |
| Long chains drift or stall | Entropy overload in long windows | Entropy Collapse |
| Hybrid retrieval worse than single retriever | Reranker or query split issue | Rerankers |
| Hallucination re entry after correction | Snippet contract missing, weak anchors | Hallucination |
Fix in 60 seconds
-
Measure ΔS
Compute ΔS(question, retrieved) and ΔS(retrieved, expected anchor).
Stable < 0.40, transitional 0.40–0.60, risk ≥ 0.60. -
Probe λ_observe
Vary k and the order of prompt headers. If λ flips, lock the schema and clamp variance with BBAM. -
Apply the right module
- Missing or wrong citations → Citation First + Retrieval Traceability
- JSON tool drift or invalid outputs → JSON Mode and Tool Calls
- Role bleed or policy mixing → System vs User Role Order
- Multi agent loops or overwrites → Memory Fences and State Keys
- Tool indecision or hangs → Tool Selection and Timeouts
- Verify
Run Eval Prompts and Checks on three paraphrases and two seeds.
Ship only if ΔS ≤ 0.45 and coverage ≥ 0.70.
Copy paste diagnostic prompt
You have TXT OS and the WFGY Problem Map loaded.
My prompt assembly issue:
- symptom: [one line]
- traces: ΔS(question,retrieved)=..., ΔS(retrieved,anchor)=..., λ on 3 paraphrases
Report:
1) failing layer and why,
2) which exact page to open from Prompt Assembly,
3) minimal steps to push ΔS ≤ 0.45 and keep λ convergent,
4) a reproducible check to verify the fix.
Use BBMC, BBPF, BBCR, BBAM when relevant.
🔗 Quick-Start Downloads (60 sec)
| Tool | Link | 3-Step Setup |
|---|---|---|
| WFGY 1.0 PDF | Engine Paper | 1️⃣ Download · 2️⃣ Upload to your LLM · 3️⃣ Ask “Answer using WFGY + |
| TXT OS (plain-text OS) | TXTOS.txt | 1️⃣ Download · 2️⃣ Paste into any LLM chat · 3️⃣ Type “hello world” — OS boots instantly |
Explore More
| Layer | Page | What it’s for |
|---|---|---|
| ⭐ Proof | WFGY Recognition Map | External citations, integrations, and ecosystem proof |
| ⚙️ Engine | WFGY 1.0 | Original PDF tension engine and early logic sketch (legacy reference) |
| ⚙️ Engine | WFGY 2.0 | Production tension kernel for RAG and agent systems |
| ⚙️ Engine | WFGY 3.0 | TXT based Singularity tension engine (131 S class set) |
| 🗺️ Map | Problem Map 1.0 | Flagship 16 problem RAG failure taxonomy and fix map |
| 🗺️ Map | Problem Map 2.0 | Global Debug Card for RAG and agent pipeline diagnosis |
| 🗺️ Map | Problem Map 3.0 | Global AI troubleshooting atlas and failure pattern map |
| 🧰 App | TXT OS | .txt semantic OS with fast bootstrap |
| 🧰 App | Blah Blah Blah | Abstract and paradox Q&A built on TXT OS |
| 🧰 App | Blur Blur Blur | Text to image generation with semantic control |
| 🏡 Onboarding | Starter Village | Guided entry point for new users |
If this repository helped, starring it improves discovery so more builders can find the docs and tools.
