Home

dotagent Wiki

One .agent/ folder. Every AI coding tool in sync. Bug registry, anti-patterns, DB hotspots, Redis keys, dependency map — every AI agent on your team reads the same canonical context, ranked by severity, every session, automatically.

Welcome. This wiki is the operational manual. For the code, see the main repo.

Start here

• Getting Started — install in 30 seconds, first sync, first commit
• Architecture — the diagram, the four memories, how everything fits
• Sources and Docs — how docs/*.md becomes the source of truth

Reference

• Commands Reference — every CLI command
• Configuration Reference — every key in .agent/config.yaml
• Memory Model — deep dive on Working / Episodic / Semantic / Personal

Features

• Auto-Dream — the agent-learning loop with mandatory rationale
• Server and RBAC — centralized event server for teams
• Multi-Project and Multi-Developer — running on N repos with N teammates

Migration & support

• Migrating from Claude-Code-Optimization — lossless migrator
• Troubleshooting — dotagent doctor, common issues, debug logging
• FAQ — answers to questions that come up

Quick reference

# install (one machine, once)
pipx install "git+https://github.com/dilawarabbas1/dotagent"

# initialize a project (each repo, once)
cd ~/code/your-project
dotagent init               # zero prompts
dotagent doctor             # self-check
git add .agent/ CLAUDE.md && git commit -m "chore: add dotagent context"

# daily flow
dotagent sync               # after editing docs/ or .agent/*.md
dotagent context            # see exactly what an AI agent sees
dotagent who --file path/to/file.py    # which dev / AI touched it
dotagent dream run          # extract learning signals

Status

v0.2.0 — launch-ready. 79+ tests passing across all phases. See the CHANGELOG.