Mysti
AGENTS.md1.2k starsChanged 6 days agoApache-2.0View on GitHub
Mysti is a VS Code extension with a host bundle, chat webview, Canvas webview, 16 provider adapters and a separate Mysti coordinator. It remains beta. Read the smallest relevant guide, then inspect its source references. package.json defines contributed settings and commands; the provider registry and capability contracts define current adapters. Plans and old release reports describe their recorded scope, not proof of current behavior. Never infer native-provider acceptance from a sample-data recording or an installed CLI. Use the Node version…
Trending
Quiet since February.about the whole repository
Who's talking about DeepMyst/Mysti
- A Hacker News story“Show HN: Mysti – Claude, Codex, and Gemini debate your code, then synthesize”Hacker News · Dec 23, 2025 · 216 points
- @matsuu.bsky.social“複数のAIエージェントを集めて解決方法の議論やブレインストーミングをさせることができるらしい。どのような議論が展開されるのかちょっと興味あるな。”Bluesky · Dec 28, 2025 · 7 likes
- @hidari.dev“github.com/DeepMyst/Mysti アイコンが、あたかもあなぅストーミングのようだ”Bluesky · Dec 28, 2025 · 2 likes
- eltociear/awesome-AI-driven-development“Add Mysti - Multi-agent AI coding assistant”GitHub · Feb 18, 2026 · 1 reactions
- jamesmurdza/awesome-ai-devtools“Add Mysti - Multi-agent AI coding assistant for VS Code”GitHub · Feb 18, 2026 · 1 reactions
The biggest 5 of 19 mentions linking to the repository that our crawl found. Each line is their own headline.
See all 19
Talked about on GitHub, Bluesky and Hacker News (19 times).
On GitHub · 15
- Feb 18, 2026Add Mysti - Multi-agent AI coding assistant — eltociear/awesome-AI-driven-development1 reaction · 1 comment
- Feb 18, 2026Add Mysti - Multi-agent AI coding assistant for VS Code — jamesmurdza/awesome-ai-devtools1 reaction
- Feb 18, 2026Add Mysti - Multi-agent AI coding assistant — Shubhamsaboo/awesome-llm-apps0 reactions · 1 comment
- Feb 18, 2026Add Mysti - Multi-agent AI coding assistant — agamm/awesome-developer-first0 reactions
- Feb 18, 2026Add Mysti - Multi-agent AI coding assistant — mahseema/awesome-ai-tools0 reactions
10 more on GitHub
- Feb 18, 2026Add Mysti - Multi-agent AI coding assistant — slavakurilyak/awesome-ai-agents0 reactions · 1 comment
- Feb 18, 2026Add Mysti - Multi-agent AI coding assistant — sorrycc/awesome-code-agents0 reactions
- Feb 18, 2026Add Mysti - Multi-agent AI coding assistant — jqueryscript/awesome-claude-code0 reactions
- Feb 18, 2026Add Mysti - Multi-agent AI coding assistant with Claude Code integration — hesreallyhim/awesome-claude-code0 reactions · 1 comment
- Feb 18, 2026Add Mysti - Multi-agent AI coding assistant for VS Code — ai-for-developers/awesome-ai-coding-tools0 reactions
- Feb 18, 2026Add Mysti - Multi-agent AI coding assistant — awesome-vibe-coding/awesome-vibe-coding0 reactions
- Feb 18, 2026Add Mysti - Multi-agent AI coding assistant — 0xWelt/Awesome-Vibe-Coding0 reactions
- Feb 18, 2026Add Mysti - Multi-agent AI coding assistant for VS Code — filipecalegario/awesome-vibe-coding0 reactions · 1 comment
- Feb 18, 2026Add Mysti - Multi-agent AI coding assistant — Hannibal046/Awesome-LLM0 reactions
- Feb 18, 2026Add Mysti - Multi-agent AI coding assistant — viatsko/awesome-vscode0 reactions
On Bluesky · 3
- Dec 28, 2025複数のAIエージェントを集めて解決方法の議論やブレインストーミングをさせることができるらしい。どのような議論が展開されるのかちょっと興味あるな。 — @matsuu.bsky.social7 likes and reposts
- Dec 28, 2025github.com/DeepMyst/Mysti アイコンが、あたかもあなぅストーミングのようだ — @hidari.dev2 likes and reposts
- Dec 28, 2025Show HN: Mysti – Claude, Codex, and Gemini debate your code, then synthesize View Article | Join the HN Conversation Summary of HN… — @hncompanion.com0 likes and reposts · 1 reply
On Hacker News · 1
- Dec 23, 2025Show HN: Mysti – Claude, Codex, and Gemini debate your code, then synthesize216 points · 178 comments
Where files are headingNew AGENTS.md files have more rules about keeping secrets out of the code, and fewer about code style.
Of the AGENTS.md files we read, the 31,774 first seen in the last six months, against the older ones: the kind of rule that gained most share, and the kind that lost most. About files like this in general, not this one, and an estimate, since a kind is a pattern in the words.
The file
We split this AGENTS.md into the rules it gives an agent
7 rules. About 1 is common, 3 are distinctive, and 3 are very specific to this project.
≈1
Common · estimate
Many files like this have a rule of this kind.
≈3
Distinctive · estimate
This author's own choices. Worth a look.
3
Very specific
Tied to their project's names, files or tools. Probably skip.
- Common
- Distinctive
- Very specific
- Very specific: line 14
Read the smallest relevant guide, then inspect its source references.
- Very specific: line 22
Use the Node version in .nvmrc and npm ci.
- Distinctive: line 40
Keep unrelated work intact.
- Common: line 44
Read the relevant provider source before changing CLI arguments, process cleanup, credentials or usage attribution.
- Distinctive: line 45
Use CONTRIBUTING.md for provider integration and security policy for vulnerability reporting.
- Very specific: line 50
Update the task guide and docs/navigation.json when adding a documented surface.
- Distinctive: line 51
Use meaningful headings, relative repository links, text alternatives for visuals, and explicit prerequisites, outcomes and limitations.
≈ “Common” is an estimate: at least 1 in 5 of the 40,246 AGENTS.md files we read have a rule of the same kind, not the same words. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.
Files
Working on Mysti
Mysti is a VS Code extension with a host bundle, chat webview, Canvas webview, 16 provider adapters and a separate Mysti coordinator. It remains beta.
Find the right document first
- Documentation map: choose a guide by task.
- Machine-readable map: paths, topics, audience and status.
- Plain-text discovery index: links to raw Markdown for remote readers.
- Architecture and developer notes: implementation boundaries.
- Production readiness: actual evidence, outstanding work and release checklist.
Read the smallest relevant guide, then inspect its source references. package.json
defines contributed settings and commands; the provider registry and capability
contracts define current adapters. Plans and old release reports describe their
recorded scope, not proof of current behavior. Never infer native-provider
acceptance from a sample-data recording or an installed CLI.
Development and validation
Use the Node version in .nvmrc and npm ci. Do not upgrade dependencies as a
workaround for a failed locked install. Start npm run watch before pressing F5.
| Change | Relevant checks |
| Documentation | npm run check:docs; inspect rendered visuals and affected walkthroughs |
| Settings or commands | npm run docs:generate, review the generated diff, then npm run check:docs |
| Application behavior | Focused regressions, npm run typecheck, npm run lint, and affected release test phases |
| Release or packaged behavior | Complete maintenance and release checklist |
Release test phases are test:release:unit, test:release:browser and
test:release:performance, in that order. Browser phases need the locked
Playwright Chromium (npm exec --no -- playwright install chromium). Keep them
separate from recording workloads. Preserve failures and skips; do not weaken
assertions, increase deadlines or add retries just to obtain a green run.
Preserve the important boundaries
- Keep unrelated work intact. Inspect
git statusbefore edits and stage only the task's files. - Bind state, permissions, async completions and cleanup to the owning panel/run.
- A streamed tool notification is not a blocking native approval. Follow approval contracts.
- Workspace settings may lower authority; they must not widen machine/user policy. Preserve default-off capability gates and untrusted-result fencing.
- Read the relevant provider source before changing CLI arguments, process cleanup, credentials or usage attribution.
- Use CONTRIBUTING.md for provider integration and security policy for vulnerability reporting.
- Publish the qualified artifact with its source identity and checksum. GitHub release, Marketplace and Open VSX publication are separate states.
Keep documentation discoverable
Update the task guide and docs/navigation.json when adding a documented surface.
Use meaningful headings, relative repository links, text alternatives for visuals,
and explicit prerequisites, outcomes and limitations. Generated references come
from the manifest; edit their generator or source rather than hand-editing output.
See documentation maintenance.
Does it work?
Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.
No reviews yet
People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.
- It worked
- Partly
- It didn't
Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.
Your trail
Files you open will collect here, so you can find them again.
Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.
Now do this for your own.
Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.

