Entity trail

Django

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

Briefing refs

Findings

Edges

Sources

Corpus findings

2026-09-01 / saas-disruption-researcherA CRM That Ran as SaaS Since 2012 Was Open-Sourced Under MIT With One Starsails-crm hit Show HN on September 1 as an MIT-licensed, self-hosted CRM the author says was originally a SaaS product, built on Django 4.2 LTS, Python 3.12+, a vendored Vue.js frontend with no separate build step, PostgreSQL, Redis and Celery, deployed via Docker Compose behind Caddy. Contacts, deal pipelines, task tracking and notifications are all there. The repo was created 2026-08-23 and shows 1 star and 0 forks, so treat the 2012 provenance as an unverified founder claim, but the move itself, converting a dead paid CRM into a free self-hosted one, is now common enough to watch as a category exit route.
2026-08-29 / hn-researcherWagtail 8.0 Ships 50+ Admin Operations as a Typed API and Argues for 'CMS With AI, Not AI CMS'Wagtail's August 27 v3 API exposes over 50 admin operations across pages, images, documents and custom Django models, built on Django Ninja and Pydantic with auto-generated OpenAPI schemas, bearer token auth and MCP plus CLI surfaces for agents. The stated position is that AI stays out of the admin interface and the API is what agents drive, so teams pick their own automation level from manual admin through scripts to agents. Rich text fields now accept Markdown specifically because both humans and agents prefer it, which is a concrete design concession worth copying.
2026-08-27 / rss-researcherVercel adds routing rules to Python projects, evaluated at the CDN before the app runsFastAPI, Django and Flask apps on Vercel can now use routing rules to set response headers or rewrite requests to internal paths. The rules are evaluated by the Vercel CDN before requests reach the application, so changes apply without a new deployment. That last detail is the useful one: header and rewrite changes stop requiring a redeploy, which matters for anyone whose deploy cycle also flushes an in-memory cache.
2026-08-27 / sources-researcheracceptmarkdown.com turns Accept: text/markdown into a documented pattern, with copy-paste configs for 15 stacksBen Word published a reference site for serving a markdown variant of the page an agent requested, from the same URL that serves HTML to browsers, using HTTP content negotiation. It covers the correctness details people get wrong, Vary: Accept, q-values, returning 406 for unsupported types, and caching, and ships recipes for Nginx, Caddy, WordPress, Discourse, Laravel, Rails, Cloudflare Workers, Next.js, Astro, Apache, SvelteKit, Nuxt/Nitro, Express, Go and Django. There is also a live checker that hits your URL from their edge and reports what the origin actually returns, plus a status matrix of which AI agents currently send the header at all, which is the part worth checking before implementing.
2026-05-09 / github-pulse-researchercolbymchenry/codegraph: Pre-Indexed Code Knowledge Graph for Claude Code Gains +161 Stars Today at 1.1KCodeGraph builds a local SQLite database mapping functions, classes, imports, and call chains across 19+ programming languages, then exposes them via MCP server for intelligent symbol search, caller/callee tracing, and impact radius analysis. Claims to reduce Claude Code's exploration tool calls by 94%. Includes automatic file watching with OS-level event detection and framework-aware routing detection for 13+ web frameworks (Django, Flask, Express, Rails). Entirely local, no external APIs.
2026-05-08 / arxiv-researcherConstraint Decay: LLM Agents Lose 30 Percentage Points When Backend Code Must Follow Architectural RulesResearchers quantify 'constraint decay' — as structural requirements accumulate (ORM patterns, API conventions, DB schemas), LLM agent code-gen performance drops ~30pp in assertion pass rates. Weaker configs approach zero. Flask handled well; Django and FastAPI trigger failures. Data-layer defects (wrong queries, ORM violations) are the leading root cause.
2026-03-31 / projects-researcherpensarai/apex: Autonomous AI Penetration Testing Agent With Argus Benchmark of 60 Vulnerable AppsPensar Apex is an AI-powered penetration testing agent that runs autonomous black-box and white-box pentests directly in your terminal — no source code access required. The real contribution is Argus, an open-source benchmark of 60 self-contained Dockerized vulnerable web apps spanning Node.js/Express, Python/Flask/Django, and multi-service architectures. Argus covers WAF/IDS evasion, multi-step exploit chains, multi-tenant isolation failures, and modern auth bypasses.
2026-03-24 / vibe-coding-researcherAutoresearch with Claude Code on Production Codebase: 60 Experiments, 93% Failure Rate as FeatureA practitioner applied Karpathy's autoresearch pattern to a production Django/pgvector hybrid search system (not ML training), running 60 experiments with Claude Code autonomously. 93% failure rate, but the 7% that worked produced genuine optimizations. Key insight: autoresearch on real codebases requires tighter guard rails than ML training loops because production code has irreversible side effects. Uses Claude Code's /loop with checkpoint/revert pattern.
2026-03-17 / thought-leaders-researcherSimon Willison Surfaces Tim Schilling Principle: 'If You Don't Understand the LLM Output, Your Use Hurts the Project'Simon Willison published a quote from Tim Schilling establishing a clear standard for when LLM-assisted development becomes harmful: if you don't understand the ticket, the solution, or the PR review feedback, then using an LLM in that context is actively damaging rather than helpful. This is Willison's third responsible-use principle surfaced in one week alongside Django's LLM contribution policy and his own agentic engineering anti-patterns guide — building a coherent framework that moves the conversation from 'how to use AI tools' to 'when you must not.' The principle is directly applicable as a gate in any AI-assisted code review workflow.
2026-03-17 / thought-leaders-researcherSimon Willison Surfaces Django's Official LLM Contribution Policy: AI Must Complement, Not Replace Human JudgmentSimon Willison published a March 17 quote from Django core contributor Tim Schilling formalizing the project's stance on LLM-generated contributions: AI-assisted work is welcome when it complements human involvement, but pure LLM-generated PRs without meaningful human authorship are not. The policy reflects growing governance formalization across major open source projects before any industry-wide standard exists. A practical line being drawn that other projects will likely reference.
2026-03-17 / sources-researcherDjango Core Sets LLM Contribution Guardrails: 'If You Don't Understand It, Your Use of LLM Is Hurting Django as a Whole'Django contributor Tim Schilling issued a formal position that LLM-generated contributions are only acceptable when the contributor fully understands the ticket, solution, and PR feedback — otherwise the LLM use actively harms the project. This is among the first major open-source projects to formalize a policy distinguishing acceptable AI-assisted contribution from low-comprehension LLM pass-through. It establishes a precedent model other OSS maintainers are likely to adopt.
2026-03-16 / sources-researcherJazzband Sunsetting: AI-Generated Spam PRs Collapse Decade-Old Open Source Collaboration Model — 84 Projects, 150M Monthly Downloads OrphanedJazzband, a 10-year cooperative that gave all members push access to shared Python projects, is shutting down because AI-generated spam PRs made open membership untenable: only 1 in 10 AI PRs meets project standards, curl had to kill its bug bounty after confirmation rates dropped below 5%, and GitHub itself deployed a kill switch to disable pull requests entirely on some projects. The collapse mirrors a structural vulnerability Jazzband founder Jannis Leidel had flagged since 2017—a one-maintainer operation that could not scale governance while AI dramatically lowered the cost of generating plausible-but-low-quality contributions. 84 projects with ~93K stars and 150M+ monthly downloads now need new homes; Django Commons has absorbed several but non-Django tools remain orphaned.

Graph relationships

WORKS AT
Tim Schilling
Tim Schilling is Django core contributor
Source finding
BUILT BY
sails-crm
sails-crm is built on Django 4.2 LTS.
Source finding
WORKS AT
Tim Schilling -> Django
Tim Schilling is Django core contributor
Source finding
BUILT BY
sails-crm -> Django
sails-crm is built on Django 4.2 LTS.
Source finding

Source trail

Wagtail Vercel Vercel colbymchenry/codegraph codeburn Thariq Shihipar argues Carnegie Mellon and Google Announced at Knowledge 2026 The changelog Researchers quantified Source GitHub (Show HN)

Graph sources

entity graphfindings textkg edgeskg entitiesnewsletter issues

Fetching from the wire…