DevTools
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.
A hub to stabilize IDE copilots and code-AI assistants without changing infra.
Every tool has its own guardrail page. Each target here has simple explanations so even new users can follow.
When to use this folder
- Unstable answers: IDE chat gives different results on the same question.
- Loops or stalls: Tool calls stop halfway or keep retrying after edits.
- Broken JSON: You expect JSON output but get plain text.
- Wrong citations: Answer looks correct but cites the wrong part of code/docs.
- Drift in long refactors: After 20–40 reasoning steps, output drifts off track.
Acceptance targets (with plain meaning)
- ΔS (semantic drift score) ≤ 0.45
How far the answer drifts from your question. Lower is better. - Coverage ≥ 0.70
How much of the correct section is included in the answer. - λ (stability factor) stays convergent
If you re-ask with small rephrasing, results stay consistent. - E_resonance flat
On long sessions, answers do not wander or change meaning.
Quick routes to per-tool pages
| Tool | Open this page |
|---|---|
| GitHub Copilot | github_copilot.md |
| Cursor | cursor.md |
| Sourcegraph Cody | sourcegraph_cody.md |
| VS Code Copilot Chat | vscode_copilot_chat.md |
| Codeium | codeium.md |
| Tabnine | tabnine.md |
| AWS CodeWhisperer | aws_codewhisperer.md |
| JetBrains AI Assistant | jetbrains_ai_assistant.md |
Map symptoms → structural fixes
| Symptom | Why it happens | Fix page |
|---|---|---|
| Wrong-meaning hits despite high similarity | Embedding captures surface form but not real meaning. | embedding-vs-semantic.md |
| Citations do not line up | Traceability layer too loose, snippet schema missing. | retrieval-traceability.md · data-contracts.md |
| Answers flip between sessions/tabs | Context not anchored, entropy builds up. | context-drift.md · entropy-collapse.md |
| JSON mode breaks, prose returned | Model leaves structured mode or prompt injection occurs. | logic-collapse.md · prompt-injection.md |
| Multi-agent or tool handoff stalls | Agents lose roles, no schema lock in exchange. | Multi-Agent_Problems.md · role-drift.md |
| Hybrid retrievers worse than single | Query parsing split or reranker mis-weighted. | pattern_query_parsing_split.md · rerankers.md |
Fix in 60 seconds
- Measure ΔS
If drift ≥ 0.60, you have unstable retrieval. - Probe λ
Re-ask with 2–3 paraphrases. If answers flip, lock schema. - Apply guardrails
- Retrieval drift → BBMC + data-contracts.md
- Reasoning collapse → BBCR bridge + BBAM + logic-collapse.md
- Dead ends → BBPF alternate paths
- Verify
Coverage ≥ 0.70, λ stable, ΔS ≤ 0.45.
Copy-paste prompt for IDE chat
I loaded TXT OS and the WFGY Problem Map.
My code-AI issue:
* symptom: \[one line]
* traces: ΔS(question,retrieved)=..., ΔS(retrieved,anchor)=..., λ states
Tell me:
1. failing layer and why,
2. the exact WFGY page to open,
3. minimal steps to push ΔS ≤ 0.45 and keep λ convergent,
4. how to verify with a reproducible test.
FAQ
Q1. What is ΔS?
It is the “semantic drift score”. Think of it as how far the answer strays from the original question. Lower numbers = more accurate.
Q2. What is λ (lambda)?
A stability check. If you slightly rephrase the same question, λ shows if the model’s answers converge or scatter.
Q3. What does E_resonance mean?
It is a long-run stability check. If E_resonance is flat, your assistant stays consistent even after many steps.
Q4. Do I need to understand BBMC/BBPF/BBCR/BBAM?
No. These are internal WFGY modules. Just know:
- BBMC = stabilizes retrieval,
- BBPF = creates fallback paths,
- BBCR = bridges reasoning collapse,
- BBAM = variance clamp for stability.
Q5. I only use one IDE plugin. Why should I care?
Because the same drift and instability happens across all copilots. Fixes here apply no matter which tool you use.
Q6. How do I test quickly?
Ask the same question 3 times with small wording changes. If answers flip, you need λ guardrails.
🔗 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 + <your question>” |
| 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.
