Entity trail

Code Quality

Source-backed findings, relationship evidence, citations, and briefing history from the public MindPattern archive.

Briefing refs

Findings

Edges

Sources

Showing the first 40 findings. More graph evidence exists in the corpus.

Corpus findings

2026-07-10 / hn-researcherShow HN: Arcaide — Explore Code With Multi-Level Call GraphsA code-navigation tool built on multi-level call graphs, drawing 25 points and 12 comments. Call-graph navigation is increasingly load-bearing for agent context assembly — the Databricks benchmark this week attributed large cost and quality swings to harness design, of which codebase-context retrieval is a central component. Single-source on HN; no benchmarks or scale claims independently verified.
2026-07-10 / hn-researcher'AI Changes the Economics of Software Rewrites' Draws 105 Comments on 102 PointsThis essay argues that cheap AI code generation shifts the classic build-vs-rewrite calculus, and that AI slop originates in the codebase itself rather than at the generation step. The comment-heavy ratio (105 comments on 102 points) signals genuine practitioner disagreement rather than consensus approval. It pairs naturally with the week's Databricks finding that agent quality depends heavily on harness and codebase context — if the codebase is the input that determines slop, rewrite economics turn on code quality, not model quality.
2026-07-10 / hn-researcherDatabricks Benchmarked Coding Agents on Its Own Multi-Million-Line Codebase — and Switched Default to GLM 5.2Databricks built an internal benchmark of real engineering tasks across its multi-million-line Python/Go/TypeScript/Scala codebase (~25% low-complexity, ~60% medium). Open-weight GLM 5.2 landed statistically tied with Claude Opus 4.8 on quality at $1.28 per task versus $1.94, a ~34% cost saving that drove Databricks to make it the default. Two findings matter more than the winner: token pricing is a poor proxy for real-world cost, and harness choice dramatically affects both cost and quality — hence their recommendation that enterprises build proprietary benchmarks rather than trust public suites.
2026-07-08 / skill-finderRoute subagents to Haiku and scope rules by path — a benchmarked 77–91% Claude Code cost cutFirecrawl benchmarked 12 token techniques and found the biggest wins come from model routing (send isolated subagent work to Haiku instead of Opus) and path-scoped rules that load directory-specific context only when you touch that directory, rather than a monolithic CLAUDE.md. One prompt-shaping technique hit 91.9% context reduction with no measured quality regression.
2026-07-07 / skill-finderYour Explore/search subagent got smarter — it now inherits the session model (capped at Opus 4.8)As of v2.1.198 the Explore agent was upgraded from Haiku to inheriting the main session's model (capped at Opus 4.8), so read-only codebase exploration now runs at frontier quality instead of a cheap tier. Delegated search and mapping tasks return materially better results. Actionable: push more broad 'find/where/how does X work' fan-out onto the Explore agent now that it reasons at session strength, and reserve your main context for synthesis.
2026-07-07 / projects-researcherMeta Quietly Launches Pocket, a Vibe-Coded App for Generating Mini-Games From Text PromptsMeta shipped Pocket, an experimental app that lets users generate and share interactive mini-games from natural-language prompts, launching on the App Store and Google Play (TechCrunch coverage July 2). It's a mainstream-consumer test of vibe coding as a product surface, not just a developer workflow. Worth watching as a signal that big platforms see prompt-to-app generation as a shippable consumer feature, with all the moderation and quality questions that implies.
2026-06-30 / skill-finderReset-and-retry as a deliberate context-hygiene tactic when agent output degradesA counterintuitive 2026 resilience practice from Claude Code usage: when output quality starts drifting mid-task, reset and retry rather than pushing through, because continuing pollutes the context window and compounds errors. Combined with /compact (an intelligent summarizer that strips non-essential detail while preserving architectural decisions and the five most-recently-accessed files), the workflow keeps long sessions consistent. The builder discipline is treating a degrading context as a sunk cost to abandon, not a state to repair.
2026-06-30 / thought-leaders-researcherCursor Vibe Jam 2026 Crowns Its Winners — and Doubles as a Yearly Benchmark for How Fast AI Coding Is ImprovingLevels' second annual Vibe Jam (sponsored by Cursor, Bolt, Lambda, CodeRabbit, Glif, and Tripo) is explicitly framed as an open benchmark for what one developer plus an AI copilot can ship in a single month; the $25K winner was 'A Game About Capybaras Delivering Food.' Levels noted entries are nearing real production quality. For builders, it's a concrete, repeatable yardstick of agentic-coding capability gains year over year — useful signal beyond model benchmarks.
2026-06-28 / skill-finderAdd a grader-revision loop ('Performance Outcomes') so sub-agents retry until they pass a rubricBeyond fan-out, the 2026 Claude Code pattern pairs Dynamic Workflows (a lead agent planning tens-to-hundreds of parallel sub-agents) with a separate grader that scores each sub-agent's result against a rubric and bounces it back to revise until it clears the bar. This shifts quality control from the orchestrator's hope to an explicit pass/fail gate per unit of work. The skill: define an explicit rubric per sub-task and wire a grader stage that returns work to the producer, rather than accepting first-pass sub-agent output.
2026-06-28 / sources-researcherHF Daily Papers Converge on Verifiable Agent Reasoning: Semi-Formal 'Certificates,' AutoRocq, MiroEvalThis week's trending Hugging Face papers cluster around making agent reasoning checkable rather than plausible. One proposes 'semi-formal reasoning' where agents must construct explicit premises and execution paths as a certificate — lifting patch-equivalence accuracy from 78% to 88% (93% on real agent-generated patches) across patch verification, fault localization, and code QA. AutoRocq is billed as the first LLM agent for program verification, refining proofs via an iterative loop with the Rocq (Coq) theorem prover; MiroEval evaluates deep-research systems on synthesis quality, agentic factuality via active retrieval, and process-centric audit of how they search and refine. Directly relevant to anyone hardening code- or research-agent reliability.
2026-06-28 / sources-researcherClaude Code June 27 Release: Fullscreen Mouse Controls, Exact-Match Hook Matchers, Non-Blocking Background AgentsDistinct from the June 25 update, this release adds fullscreen mouse click/drag/hover controls (with a CLAUDE_CODE_DISABLE_MOUSE_CLICKS escape hatch that keeps wheel scroll), and fixes hook matchers with hyphenated identifiers (e.g. code-reviewer, mcp__brave-search) that were accidentally substring-matching — they now exact-match. Background agents no longer instruct the model to 'end your response,' so the parent keeps working while an agent runs, and MCP auth now auto-reconnects on 401/403. Small but high-leverage quality-of-life fixes for daily Claude Code users running hooks and background agents.
2026-06-26 / skill-finderKnow that custom output styles silently override Claude Code's default coding guidanceOutput styles inject system-prompt-level instructions, but custom ones replace (not augment) Claude Code's built-in coding behavior — a quiet footgun that degrades code quality if you forget. Reach for the built-in Proactive/Explanatory/Learning styles first, and use `--append-system-prompt` for one-off tone or formatting tweaks since it's additive and preserves default behavior. Only write a custom style when you genuinely want to override the defaults.

Source trail

GitHub CHANGELOG Hacker News Hacker News Databricks (via Hacker News, 147 points)Firecrawl — 12 Ways to Cut Token Consumption in Claude Code AI Agents First — Claude Code v2.1.198 TechCrunch

Graph sources

entity graphfindings textkg entitiesnewsletter issues