balanced7/akashic-aurora
Memory for AI agents that learns what actually helped: lessons injected at the moment of action, credited by outcome, challenged by counter-evidence. Local-first, multi-agent, test-gated.
ARCHETYPE
YOLO Cowboy
Permissions? Never heard of 'em. Ships at the speed of `--dangerously-skip-permissions`.
Copy this rig
# review before running: this installs third-party code $ claude mcp add akashic-aurora -- py ai_setup_mcp.py $ npx degit balanced7/akashic-aurora/.claude ./rig-akashic-aurora # inspect, then merge into .claude/
MCP servers are added to Claude Code at local scope; env vars are shown as YOUR_… placeholders — we never store values. Files are fetched with degit into a separate folder so you can review before merging.
$ claude mcp add akashic-aurora -- py ai_setup_mcp.py $ npx degit balanced7/akashic-aurora/.agents/skills/akashic-memory .claude/skills/akashic-memory $ npx degit balanced7/akashic-aurora/.agents/skills/plan-with-the-corpus .claude/skills/plan-with-the-corpus $ npx degit balanced7/akashic-aurora/.agents/skills/root-cause-before-fixes .claude/skills/root-cause-before-fixes $ npx degit balanced7/akashic-aurora/.agents/skills/verified-done .claude/skills/verified-done $ npx degit balanced7/akashic-aurora/.claude/skills/deconstruct .claude/skills/deconstruct $ npx degit balanced7/akashic-aurora/.claude/skills/reverse-engineering .claude/skills/reverse-engineering $ npx degit balanced7/akashic-aurora/.claude/skills/shader-craft .claude/skills/shader-craft $ npx degit balanced7/akashic-aurora/research/in-flight/video-primitives-2026-10-12/skills/video-work .claude/skills/video-work $ npx degit balanced7/akashic-aurora/research/in-flight/video-primitives-2026-10-12/skills/vod-review .claude/skills/vod-review
Merge into .claude/settings.json (project) or ~/.claude/settings.json (user). Hook commands reference scripts in the source repo — copy those too.
{
"hooks": {
"PreToolUse": [
{
"matcher": "Bash|PowerShell|Read|Edit|Write|NotebookEdit|Glob|Grep|Task|WebFetch|WebSearch",
"hooks": [
{
"type": "command",
"command": "pyw agent/harness/hooks/claude_trace.py"
}
]
}
]
}
} MCP servers (1)
| server | source | est. tokens |
|---|---|---|
| py · "akashic-aurora" | binary | 2.5k |
Skills (9)
akashic-memoryplan-with-the-corpusroot-cause-before-fixesverified-donedeconstructreverse-engineeringshader-craftvideo-workvod-review
Hooks (3)
| event | matcher | runs |
|---|---|---|
| PreToolUse | Bash|PowerShell|Read|Edit|Write|NotebookEdit|Glob|Grep|Task|WebFetch|WebSearch | pyw agent/harness/hooks/claude_trace.py |
| PostToolUse | Bash|Edit|Write|NotebookEdit | pyw agent/harness/hooks/claude_posttooluse.py |
| Stop | * | pyw agent/harness/hooks/claude_stop.py |
Cursor rules (1)
akashic-memory · always
Permissions
deny (0)
—
ask (0)
—
allow (50)
Bash(py agent_cli.py boot*)
Bash(py agent_cli.py delta*)
Bash(py agent_cli.py discover*)
Bash(py agent_cli.py recall*)
Bash(py agent_cli.py notes*)
Bash(py agent_cli.py status*)
Bash(py agent_cli.py stats*)
Bash(py agent_cli.py injections*)
Bash(py agent_cli.py harnesses*)
Bash(py agent_cli.py triage*)
Bash(py agent_cli.py story*)
Bash(py agent_cli.py events*)
Bash(py agent_cli.py doctor*)
Bash(py agent_cli.py promoted*)
Bash(py agent_cli.py lookback*)
Bash(py agent_cli.py knowledge-map*)
Bash(py agent_cli.py flow*)
Bash(py agent_cli.py locks*)
Bash(py agent_cli.py task list*)
Bash(py agent_cli.py bifrost-sync*)
mcp__akashic-aurora__boot
mcp__akashic-aurora__recall
mcp__akashic-aurora__recall_at
mcp__akashic-aurora__recall_feedback
mcp__akashic-aurora__notes
mcp__akashic-aurora__promoted
mcp__akashic-aurora__locks
mcp__akashic-aurora__stats
mcp__akashic-aurora__events
mcp__akashic-aurora__knowledge_map
mcp__akashic-aurora__bifrost_sync
mcp__akashic-aurora__bifrost_presence
mcp__akashic-aurora__injections
Bash(py agent_cli.py learn*)
Bash(py agent_cli.py note*)
Bash(py agent_cli.py handoff*)
Bash(py agent_cli.py lock*)
Bash(py agent_cli.py unlock*)
Bash(py agent_cli.py log*)
Bash(py agent_cli.py bifrost-send*)
Bash(py agent_cli.py bifrost-ack*)
mcp__akashic-aurora__learn
mcp__akashic-aurora__note
mcp__akashic-aurora__handoff
mcp__akashic-aurora__lock
mcp__akashic-aurora__log
mcp__akashic-aurora__bifrost_send
Edit(research/reviewed/**)
Bash(py scripts/bifrost_wake.py*)
Bash(py agent_cli.py unwedge*)
Similar rigs
DesiRichDev/Claude-Cowork-Linkedin-X
Social content agent for LinkedIn and X. Drop in your brand config, research signals, write platform-native posts, review quality, push to Supabase. Built on Claude Code.
Pragmatist 505 tok · · Q0
Fixed-Seed/game-changer
The AI game-modding toolkit from Fixed Seed: skills, the fsgc CLI, a FixedSeed MCP server and a shared knowledge base so any coding agent can mod almost any PC game you own.
Skill Collector 4.7k tok · · Q15
rehan-remade/universal-modder
Point Claude at any game. Skills, tools and the fal MCP that let Claude Code mod almost any PC game you own: recon, reverse engineering, fal-generated art/3D/audio, in-game testing, showcase videos.
Skill Collector 4.6k tok · · Q15
blueclover22/mw_agent_kit
Spec-driven development workflow for Claude Code — built around an approved design doc, where each stage waits for your approval.
Orchestrator 3.0k tok · · Q3