pingLABSCO SUMMARY
It solves the moment-of-edit problem: the rule shows up where the change is being made, not in a file at the top of the repo.
A top-level instruction file is read once and forgotten by the time the agent is deep in a function. Putting the note at a mirror path means it is reachable from the file itself, with no search step to get wrong. The drift checking is what keeps that honest — a note whose source has moved on is flagged rather than trusted, which is the failure mode every hand-written memory system eventually hits. The cost is process: this expects worktrees, gates and a closeout step, so it fits teams willing to adopt a workflow rather than just install a server.
What it is
A Python server that keeps versioned Markdown notes beside your code at deterministic mirror paths, verifies them against source commits before use, and updates them only after approved work lands.
What you get
- Path-addressed notes: a source file's context lives at a predictable mirror path, so an agent holding the file reaches the right note without searching or ranking
- context_packet — the orienting read at the start of a task
- memory_quality_check and drift_check — notes tested against source commits and fingerprints before they are believed
- memory_init and memory_baseline_status — a repository brought under memory management, and its baseline state reported
- resolve_context, for pulling the notes relevant to what is being changed
- Isolated dual worktrees through worktree_start, worktree_status, worktree_cleanup and worktree_abandon, so memory and code land together or not at all
- gate_decide and gate_list, for the approval points in the lifecycle
- dispatch_agent and the parent/child messaging tools, for coordinating more than one agent
- Repo-owned behaviour in system files — path rules, tools, coding guidelines, branch policy, reporting shape — so the same rules load whichever harness you use
- Starter packages for Claude Code, Codex, Cursor, Antigravity, VS Code Copilot, Hermes, Pi.dev and OpenClaw
Requirements
uv for uvx, or pip, with Python 3.11 or higher, and Git configured with a user name and email. Memory lives in the repository under ar-memory by default; teams wanting separate memory repositories use the external layout instead. No account and no key — everything is files and Git.
Setup effort
One command — uvx agents-remember-mcp@latest --config <YOUR_PATH_TO_AGENTS_REMEMBER_SETTINGS>
USE CASES
01
Making sure an agent knows the rule before it edits the file
The note sits at the file's mirror path, so it is read alongside the source rather than hoped for.
02
Knowing whether the memory is still true
Notes are drift-checked against the commits they describe, so stale context is flagged instead of quietly used.
03
Starting a task with the agent actually oriented
One orienting read at task start, followed by a memory health check, before any change is proposed.
04
Keeping code and its documentation in step
Work happens in paired worktrees and lands together, so memory does not drift behind the merge.
05
Bringing a repository under memory for the first time
Initialise the memory layer and read back its baseline before relying on any of it.
06
Running the same project rules across different agent tools
The system files live in the repo, and the starter packages carry what each harness needs to load them.
48 tools
Listed in project docs. Your client may show a different list.
server_inforesolve_contextcontext_packetdispatch_agentmessage_parentmessage_childretire_childrename_childrename_selflifecycle_gategate_decidegate_listruntime_installskills_installmemory_initdrift_checkmemory_quality_checkcitation_fixroute_index_refreshread_ar_filesmemory_baseline_statusmemory_baseline_adoptmemory_carryover_planmemory_carryover_applyworktree_startworktree_attachworktree_statusworktree_closeout_previewworktree_closeout_applyworktree_integrateworktree_cleanuplifecycle_finalize_tasktask_doctask_reopenprovider_statusprovider_diagnosticsprovider_watchersgrepai_searchgrepai_tracecgc_symbol_searchcgc_callerscgc_calleescgc_dependenciescgc_complexitycgc_visualizecodex_benchmark_preparecodex_benchmark_runBEFORE YOU INSTALL
What to know first
Tools you get
48 tools named in the project's documentation.
Requires
uvx on PATH
INSTALLATION
Connect your client
Server setup
TERMINAL
uvx agents-remember-mcp@latest --config <YOUR_PATH_TO_AGENTS_REMEMBER_SETTINGS>MCP Clients
Settings file
Every route on this page needs --config pointing at this settings file. The starter package for your harness (under docs/install in the repository) ships one at mcp/agents-remember-settings.json; or save a minimal one like this outside the ar-coordination folder:
SETTINGS FILE · JSON
{
"version": 1,
"coordinationRoot": "<YOUR_COORDINATION_ROOT>",
"workspaceRoot": "<YOUR_WORKSPACE_ROOT>",
"repositories": {
"<YOUR_REPO_NAME>": {}
},
"providers": {
"codegraphcontext-code": {},
"grepai-memory": {}
}
}<YOUR_REPO_NAME> is absolute paths, and your repository's folder name. The two providers entries are the optional Docker-backed providers, which need Docker running. Every field is described in docs/reference/settings-json.md in the repository.
OpenClaw
OPENCLAW · JSON
{
"mcp": {
"servers": {
"agents-remember": {
"command": "uvx",
"args": [
"--refresh-package",
"agents-remember-mcp",
"agents-remember-mcp@latest",
"--config",
"<YOUR_PROJECTS_FOLDER>/.openclaw/mcp/agents-remember-settings.json"
]
}
}
}
}Hermes
HERMES · TERMINAL
uvx --refresh-package agents-remember-mcp agents-remember-mcp@latest --config <YOUR_PROJECTS_FOLDER>/.hermes/mcp/agents-remember-settings.json
