🧭 Quick Return to Map
You are in a sub-page of DevTools_CodeAI.
To reorient, go back here:
- DevTools_CodeAI — AI-assisted coding and developer productivity
- 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 desk within a ward.
If you need the full triage and all prescriptions, return to the Emergency Room lobby.
Use this page when Copilot autocompletes the wrong API, flips between two implementations, breaks tests after a long chat, or injects code that does not match the project’s version and policy. The fixes here route back to WFGY pages with measurable targets so you can verify without changing infra.
- Visual map and recovery: RAG Architecture & Recovery
- End to end retrieval knobs for code search and docs: Retrieval Playbook
- Why this snippet and how to audit it: Retrieval Traceability
- Ordering and ranking control: Rerankers
- Similarity is high yet meaning is wrong: Embedding ≠ Semantic
- Chunk and citation boundaries for code and docs: Hallucination
- Long chats that drift: Context Drift, Entropy Collapse
- Structural collapse and recovery: Logic Collapse
- Prompt injection in code assistants and terminals: Prompt Injection
- Build and deploy order problems: Bootstrap Ordering, Deployment Deadlock, Pre-deploy Collapse
- Snippet and citation schema for code: Data Contracts
- ΔS(question, retrieved) ≤ 0.45 for the cited file or doc section
- Coverage ≥ 0.70 to the target function or spec anchor
- λ remains convergent across three paraphrases and two seeds
- E_resonance flat on long windows while editing
-
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 Change k for code search candidates and toggle deterministic reranker. If λ flips when prompt headers reorder, lock the schema and clamp with BBAM.
-
Apply the right module
- Retrieval drift in file or API lookup → BBMC with Data Contracts
- Reasoning collapse in long refactors → BBCR bridge plus BBAM, verify with Logic Collapse
- Dead ends during multi step edits → BBPF alternate paths
- Hybrid search worse than single → pattern check Query Parsing Split and enforce Rerankers
-
Hallucinated or wrong version APIs Lock SDK version and import surface in the prompt context. Require file anchors and commit hash in the schema. Open: Retrieval Traceability, Data Contracts, Embedding ≠ Semantic.
-
Refactor touches one file but not callers Split the plan into independent paths and rejoin with a bridge step. Open: Context Drift, Entropy Collapse.
-
Autocomplete uses private secrets or copies unsafe snippets Enforce a schema that forbids secrets and requires license note for pasted code. Open: Prompt Injection, Data Contracts.
-
Build fails after first apply The tool fired before dependencies were ready or used a stale index. Open: Bootstrap Ordering, Pre-deploy Collapse.
-
Hybrid search worse than ripgrep or a single retriever HyDE plus BM25 diverged and split the query. Open: Pattern: Query Parsing Split, Rerankers.
- Index state proves readiness. Cache key includes language server version and analyzer settings.
- Code search has a single metric per run. Do not mix analyzers within one fix loop.
- Prompts carry explicit anchors: file path, function name, commit hash, license flag.
- Every edit records citations and ΔS, plus λ per step.
- Regression gate runs tests and requires coverage ≥ 0.70 and ΔS ≤ 0.45 before commit.
You have TXTOS and the WFGY Problem Map loaded.
My Copilot issue:
- symptom: [one line]
- project: {repo}@{commit}
- traces: ΔS(question,retrieved)=..., ΔS(retrieved,anchor)=..., λ across 3 paraphrases
Do:
1) Identify failing layer and why.
2) Name the exact WFGY page to open from the repo.
3) Give minimal steps to push ΔS ≤ 0.45 and keep λ convergent.
4) Return a short, auditable plan with file anchors and a test to verify.
Use BBMC, BBPF, BBCR, BBAM when relevant.-
ΔS stays ≥ 0.60 after enforcing anchors and reranking Rebuild the code index with consistent analyzer, then re test with a small gold set. See: Embedding ≠ Semantic, Retrieval Playbook.
-
Suggestions flip between runs with identical context Lock prompt headers and clamp variance. See: Context Drift, Data Contracts.
| 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 |
| 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.
