Sourcegraph Cody: Guardrails and Fix Patterns

🧭 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.

A field guide to stabilize Cody when it mixes repo search, embeddings, cross-repo context, refactors, and tests. Use this to localize the failure, then jump to the exact WFGY fix page with measurable acceptance.

Open these first

Visual map and recovery: RAG Architecture & Recovery
End to end retrieval knobs: Retrieval Playbook
Why this snippet and how to audit: Retrieval Traceability
Ordering and ranking control: Rerankers
High similarity yet wrong meaning: Embedding ≠ Semantic
Chunk and citation boundaries: Hallucination
Long chats that drift: Context Drift, Entropy Collapse
Structural collapse and recovery: Logic Collapse
Multi-agent conflicts and tool handoffs: Multi-Agent Problems
Boot order problems in IDE flows: Bootstrap Ordering, Deployment Deadlock, Pre-deploy Collapse
Snippet and citation schema: Data Contracts

Core acceptance

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 stays flat across long edit plans

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 Switch context provider between embeddings and code search. Vary k as 5, 10, 20. If ΔS stays flat and high, suspect metric or index mismatch. If λ flips with header reorder, lock the schema and clamp with BBAM.
Apply the module

Retrieval drift in code or doc lookup → BBMC plus Data Contracts
Reasoning collapse during multi step refactors → BBCR bridge plus BBAM, verify with Logic Collapse
Dead ends in long edit plans → BBPF alternate paths
Hybrid search worse than single → Pattern Query Parsing Split and Rerankers

Typical Cody breakpoints and the right fix

Out of date code graph or precise index Symbol lookups land on old definitions. Warm up precise index, verify commit at HEAD, fence first run. Open: Bootstrap Ordering, Retrieval Traceability
Monorepo root mismatch Context comes from sibling packages. Lock anchors to repo@commit and subdir. Open: Data Contracts
Permissions filter the context Cody falls back to public snippets. Require citation fields to echo repo, commit_sha, file_path. Open: Retrieval Traceability
High similarity yet wrong symbol Embeddings hit near neighbors. Rebuild with explicit metric and normalization; fix store fragmentation. Open: Embedding ≠ Semantic, Vectorstore Fragmentation
Two stage queries drift HyDE plus code search split the intent. Pin the pair and use deterministic rerank. Open: Query Parsing Split, Rerankers
Generated tests reference phantom helpers Enforce cite-then-explain with file anchors and commit SHA before generating tests. Open: Data Contracts
Long chat drift after 25–40 steps Plan degrades or re-introduces removed code. Split plan and rejoin with a BBCR bridge. Open: Context Drift, Entropy Collapse

IDE checklist for Cody

Warm-up fence checks INDEX_HASH, precise index present, credentials valid.
Single retrieval metric per run. Do not mix analyzers while fixing one bug.
Prompts carry anchors: repo@commit, file_path, symbol, line_start, line_end, snippet_id.
Log per step: ΔS, λ state, coverage. Alert when ΔS ≥ 0.60 or λ diverges.
Regression gate requires tests pass, coverage ≥ 0.70, ΔS ≤ 0.45, same diff twice.

Minimal schema you should capture

{ repo, commit_sha, file_path, symbol, line_start, line_end, snippet_id, tokens, ΔS, λ_state } Require cite-then-explain. Forbid cross file reuse without a new citation.

Deep diagnostics

Three paraphrase probe Ask the same change three ways. If λ flips with harmless header reorder, clamp with BBAM and lock the schema.
Anchor triangulation Compare ΔS for the intended file vs a decoy file or sibling package. If close, re-chunk and normalize embeddings. See: Retrieval Playbook, Embedding ≠ Semantic
Plan length audit If entropy rises after 25–40 steps, split the plan, re-join with a BBCR bridge. See: Entropy Collapse
Live instability Add probes and backoff guards in Cody tasks. See: Live Monitoring for RAG, Debug Playbook

Escalate and structural fixes

Index or metric mismatch persists Rebuild with consistent analyzer, verify with a small gold set and pinned rerankers. Open: Rerankers, Chunking Checklist
Outputs flip between identical runs Lock prompt headers, clamp variance, enforce deterministic patch flow. Open: Logic Collapse

Copy-paste prompt for Cody chat

You have TXTOS and the WFGY Problem Map loaded.

My Cody issue:
- symptom: [one line]
- anchors: repo={name}, commit={sha}, file={path}, lines={a..b}
- traces: ΔS(question,retrieved)=..., ΔS(retrieved,anchor)=..., λ across 3 paraphrases

Tell me:
1) the failing layer and why,
2) the exact WFGY page to open,
3) minimal steps to push ΔS ≤ 0.45 and keep λ convergent,
4) a reproducible test to verify the fix.
When relevant, use BBMC, BBPF, BBCR, BBAM. Keep it auditable and short.

🔗 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.

12 KiB Raw Permalink Blame History Unescape Escape