feed · watching · stats · help with your tokens
mycelium-hq/ai-brain-starter
AI Brain Starter is a verification harness for Claude Code that provides persistent memory, deterministic guards, and session lifecycle management to prevent context corruption.
It solves the problem of AI agents forgetting context between sessions by providing a persistent memory harness with deterministic guards and session rituals for professionals who need reliable context management. It uses a local Markdown vault as ground truth, deterministic pre-write hooks to block bad edits, and a bi-temporal rule lineage to detect drift and manage memory.
Similar
- JustVugg/colibri
colibri is a C inference engine that runs very large open mixture-of-experts models on the machine you already have.
- deepseek-ai/dsh-superagent-prototype
dsh-superagent is a research prototype for a proactive coding agent that works on a single GitHub repository using DeepSeek Harness.
- anthropics/FlashProve
FlashProve is a research prototype for zero-knowledge proofs of AI inference that allows providers to prove model identity without revealing weights.
- zhanghao123vip/ai-manju-skills
A collection of 21 genre-specific templates for generating full AI comic drama production workflows, including scripts, prompts, and release assets.
- pgEdge/ai-dba-workbench
A PostgreSQL monitoring system that combines traditional and AI-based monitoring for instances v14+ including Supabase and Amazon RDS.
- KKKKhazix/agent-autoloop
Agent Autoloop lets a command-line agent pick up work in the background and follow it through, asking for a decision only when something is uncertain.
- NVIDIA/auto-ontology
An open context layer that turns database metadata into an ontology, so AI agents answer questions about enterprise data with fewer wrong answers.
- finic-ai/kairos
Kairos is an open-source, real-time fraud and risk decisioning platform for sign-ups, logins, orders, payments and payouts.
sources
Claims below include README evidence. A matching quote or worker agreement does not establish that a claim is correct. Category majority among 2 workers. Version 1.
AI Brain Starter is a verification harness for Claude Code that provides persistent memory, deterministic guards, and session lifecycle management to prevent context corruption.
A verification harness around your AI agent. So memory compounds instead of corrupts.
Quote matched the README or GitHub description at extraction. Same claim and quote from 1 worker(s). Interpretation not reviewed by a maintainer.
It solves the problem of AI agents forgetting context between sessions by providing a persistent memory harness with deterministic guards and session rituals for professionals who need reliable context management.
You have a smart assistant that forgets you every morning. This is the harness that gives it a memory: notes on your disk that the assistant reads before answering you, deterministic guards that block bad edits before they land, and a session-end ritual that files what mattered.
Quote matched the README or GitHub description at extraction. Same claim and quote from 1 worker(s). Interpretation not reviewed by a maintainer.
It uses a local Markdown vault as ground truth, deterministic pre-write hooks to block bad edits, and a bi-temporal rule lineage to detect drift and manage memory.
Vault as ground truth. Markdown on disk. Decisions, patterns, context. The model reads from the vault, not from chat history.
Quote matched the README or GitHub description at extraction. Same claim and quote from 1 worker(s). Interpretation not reviewed by a maintainer.
Stores decisions, patterns, and context in a local Markdown vault that the model reads from instead of chat history.
Vault as ground truth. Markdown on disk. Decisions, patterns, context.
Quote matched the README or GitHub description at extraction. Same claim and quote from 1 worker(s). Interpretation not reviewed by a maintainer.
Uses pre-write hooks to deterministically block bad edits, such as missing frontmatter fields or raw git add commands on large vaults.
Hooks as deterministic guards. Pre-write hooks block bad edits before they land.
Quote matched the README or GitHub description at extraction. Same claim and quote from 1 worker(s). Interpretation not reviewed by a maintainer.
Tracks rule lineage with two timestamps to detect conflicts and scan for cumulative semantic shift in high-edit files.
Bi-temporal rule lineage. Every codified rule carries two clocks: when it was written, and when it was last verified.
Quote matched the README or GitHub description at extraction. Same claim and quote from 1 worker(s). Interpretation not reviewed by a maintainer.
Manages session lifecycle by loading context at start and cascading decisions, insights, and to-dos to files at end.
Session lifecycle. Each session start loads context. Each session end cascades decisions, insights, and to-dos to the right files.
Quote matched the README or GitHub description at extraction. Same claim and quote from 1 worker(s). Interpretation not reviewed by a maintainer.
Runs on a standard Claude subscription without requiring a vector database, fine-tuning, or per-seat licenses.
Free to install. Free to run. Your Claude subscription is the inference. No vector database, no fine-tune, no per-seat license.
Quote matched the README or GitHub description at extraction. Same claim and quote from 1 worker(s). Interpretation not reviewed by a maintainer.
Includes an advisory panel of custom voices that reviews drafts and pushes back on weak thinking.
The advisory panel (Naval, Brene Brown, Hormozi, Buffett, plus the custom voices you build) meets your draft and pushes back where your thinking is soft.
Quote matched the README or GitHub description at extraction. Same claim and quote from 1 worker(s). Interpretation not reviewed by a maintainer.
Processes meeting transcripts to file them, tag participants, extract action items, and route them to owners by role.
A one-hour meeting ends. You drop the transcript in. The system files it, tags participants, extracts five action items, routes each one to the right owner by role
Quote matched the README or GitHub description at extraction. Same claim and quote from 1 worker(s). Interpretation not reviewed by a maintainer.
Checks writing continuity against previous chapters and scores voice match against the user's own corpus.
Claude reads the new scene against every previous chapter in the vault, flags a continuity break in chapter three, and scores the voice match against your own corpus.
Quote matched the README or GitHub description at extraction. Same claim and quote from 1 worker(s). Interpretation not reviewed by a maintainer.
Runs a weekly ritual that surfaces all open loops from the last fourteen days in a single view.
The weekly ritual runs. Every open loop from the last fourteen days surfaces in a single view, across the whole life
Quote matched the README or GitHub description at extraction. Same claim and quote from 1 worker(s). Interpretation not reviewed by a maintainer.
Supports team integration by syncing a team folder bidirectionally with a separate shared vault without leaking personal data.
Inside it, a team folder syncs bidirectionally with a separate team vault that your co-founders, operators, or contractors share.
Quote matched the README or GitHub description at extraction. Same claim and quote from 1 worker(s). Interpretation not reviewed by a maintainer.
Stores decisions, patterns, and context in a markdown vault on disk that the AI reads instead of relying on chat history.
Vault as ground truth. Markdown on disk. Decisions, patterns, context. The model reads from the vault, not from chat history.
Quote matched the README or GitHub description at extraction. Same claim and quote from 1 worker(s). Interpretation not reviewed by a maintainer.
Uses pre-write hooks to block bad edits, enforce frontmatter requirements, and validate calendar events before they are saved.
Hooks as deterministic guards. Pre-write hooks block bad edits before they land. Em dashes in external prose. Frontmatter without required fields.
Quote matched the README or GitHub description at extraction. Same claim and quote from 1 worker(s). Interpretation not reviewed by a maintainer.
Tracks when rules were written and verified to detect conflicts and promote hardened patterns while demoting stale ones.
Bi-temporal rule lineage. Every codified rule carries two clocks: when it was written, and when it was last verified. The resolver detects conflicts.
Quote matched the README or GitHub description at extraction. Same claim and quote from 1 worker(s). Interpretation not reviewed by a maintainer.
Automatically loads context at session start and cascades decisions, insights, and to-dos to files at session end.
Session lifecycle. Each session start loads context. Each session end cascades decisions, insights, and to-dos to the right files.
Quote matched the README or GitHub description at extraction. Same claim and quote from 1 worker(s). Interpretation not reviewed by a maintainer.
Runs locally using the user's existing Claude subscription without requiring vector databases, fine-tuning, or per-seat licensing.
Free to install. Free to run. Your Claude subscription is the inference. No vector database, no fine-tune, no per-seat license.
Quote matched the README or GitHub description at extraction. Same claim and quote from 1 worker(s). Interpretation not reviewed by a maintainer.
Integrates an advisory panel of custom voices to critique drafts and improve thinking clarity before the user finalizes content.
The advisory panel (Naval, Brene Brown, Hormozi, Buffett, plus the custom voices you build) meets your draft and pushes back where your thinking is soft.
Quote matched the README or GitHub description at extraction. Same claim and quote from 1 worker(s). Interpretation not reviewed by a maintainer.
Automatically processes meeting transcripts to tag participants, extract action items, route tasks by role, and draft follow-up messages.
The system files it, tags participants, extracts five action items, routes each one to the right owner by role, and drafts the follow-up message for you to review.
Quote matched the README or GitHub description at extraction. Same claim and quote from 1 worker(s). Interpretation not reviewed by a maintainer.
Enforces required fields for tasks to prevent incomplete entries, ensuring data quality before saving.
A hook blocks the save: missing the four required fields (source, location, shape, channel). You rewrite the task in ninety seconds.
Quote matched the README or GitHub description at extraction. Same claim and quote from 1 worker(s). Interpretation not reviewed by a maintainer.
Runs a weekly ritual to surface all open loops from the last two weeks across business, personal, and professional contexts.
The weekly ritual runs. Every open loop from the last fourteen days surfaces in a single view, across the whole life: the company, the book, the consulting, the people you care about.
Quote matched the README or GitHub description at extraction. Same claim and quote from 1 worker(s). Interpretation not reviewed by a maintainer.
Includes a marketplace of skills and MCP servers optimized for specific team workflows, stacks, and decision logic.
A marketplace of skills and MCP servers custom-built for your team's exact workflows, optimized to your stack, your voice, your decision logic.
Quote matched the README or GitHub description at extraction. Same claim and quote from 1 worker(s). Interpretation not reviewed by a maintainer.
Ensures local installation where vault data never leaves the machine, with optional email being the only data sent to servers.
Local install. Your vault data never leaves your machine. The email — if you choose to give one — is the only thing that touches our servers.
Quote matched the README or GitHub description at extraction. Same claim and quote from 1 worker(s). Interpretation not reviewed by a maintainer.
Limitation: Requires Claude Code version 2.1.133 or newer to run.
Requires Claude Code 2.1.133 or newer. Bootstrap will refuse to run on older versions and tell you the upgrade command.
Quote matched the README or GitHub description at extraction. Same claim and quote from 1 worker(s). Interpretation not reviewed by a maintainer.
Limitation: On Windows, only obsidian-skills and two MCP servers are installed; seven vendor skill-packs are not yet ported.
On Windows, bootstrap.ps1 installs only obsidian-skills plus the two MCP servers below; the seven vendor skill-packs are not ported yet.
Quote matched the README or GitHub description at extraction. Same claim and quote from 1 worker(s). Interpretation not reviewed by a maintainer.
Limitation: Requires Claude Code version 2.1.133 or newer; older versions will refuse to run the bootstrap script.
Requires Claude Code 2.1.133 or newer. Bootstrap will refuse to run on older versions and tell you the upgrade command.
Quote matched the README or GitHub description at extraction. Same claim and quote from 1 worker(s). Interpretation not reviewed by a maintainer.
Limitation: Windows installation only includes obsidian-skills and two MCP servers; seven vendor skill-packs are not yet ported to Windows.
On Windows, bootstrap.ps1 installs only obsidian-skills plus the two MCP servers below; the seven vendor skill-packs are not ported yet.
Quote matched the README or GitHub description at extraction. Same claim and quote from 1 worker(s). Interpretation not reviewed by a maintainer.
Limitation: Preserves existing custom configurations using setdefault semantics and never overwrites them during installation.
Existing custom marketplaces, plugins, hooks, env vars, and permissions are preserved (setdefault semantics, never overwrites).
Quote matched the README or GitHub description at extraction. Same claim and quote from 1 worker(s). Interpretation not reviewed by a maintainer.
Limitation: Vendor skill-packs on Windows must be added manually after installation by running specific commands in Claude Code.
To add any of them manually after install, paste this into Claude Code (substituting names from the table above): /plugin marketplace add /
Quote matched the README or GitHub description at extraction. Same claim and quote from 1 worker(s). Interpretation not reviewed by a maintainer.
Limitation: Full installation contract and post-install details are located in the assistant guide at the end of the README section.
The full contract, and what comes after the install, is in the assistant guide at the end of this section.
Quote matched the README or GitHub description at extraction. Same claim and quote from 1 worker(s). Interpretation not reviewed by a maintainer.
Limitation: If IT approval is needed, the system generates a copy-pasteable request in English and Spanish naming required packages and reasons.
If anything needs IT, it prints a copy-pasteable request naming the exact packages and why, in English and Spanish together, so one message covers a mixed team.
Quote matched the README or GitHub description at extraction. Same claim and quote from 1 worker(s). Interpretation not reviewed by a maintainer.
ai-ml