$ loading_
Prevents stale writes on shared files so agents do not silently overwrite each other.
Copy the install command and let the AI configure it · recommended for beginners
Please install the "Agent Coherence — Stale Write Guard (FS)" MCP server from askskill: Run: claude mcp add 'io-github-cohexa-ai-stale-write-guard-fs' -- uvx agent-coherence
Before modifying the shared config file, use Agent Coherence — Stale Write Guard (FS) to verify the current version; if another agent has updated the file, deny this write and tell me to reread and merge changes first.
The tool blocks the write when it detects the file version has changed, preventing old content from overwriting newer changes.
Two agents are editing the same documentation file. At write time, use this tool to validate that the file is still current; if not, stop the write and return a conflict warning.
The result indicates whether the write is denied or allowed, helping the team catch and handle concurrent editing conflicts.
When an automation workflow writes back to a shared results file, first use Stale Write Guard to check whether the file changed since it was read; if it did, do not write and log the reason.
The tool rejects stale writes when detected, reducing the risk of data loss in automation workflows.
When developers or automation agents work on the same file at the same time, this tool can block writes based on outdated state to avoid silent overwrites. It makes conflicts visible and encourages resyncing before changes are applied.
When updating shared configs, scripts, or other files, it serves as a coherence guard that rejects stale writes. It is useful in engineering or DevOps situations where concurrent edits are sensitive.
It is a coherence guard for shared files that rejects stale writes. This prevents multiple agents from silently overwriting each other when operating on the same file.
Based on the provided information, its core capability is denying stale writes and preventing silent overwrites. Whether it also supports automatic merge or conflict resolution is not stated; see the source repository.
No installation guide or prerequisite details were provided here. For setup steps, runtime dependencies, and configuration requirements, see the source repository.
Coordinate multiple coding agents safely within the same repository.
Safely edit files in MCP clients with hash checks against stale writes.
Maintain append-only shared memory and review-based acceptance for collaborative work.
Adds human approval and policy checks to risky AI file operations.
Coordinate multiple AI agents in one Git repo and prevent conflicts early.
Coordinate AI coding agents with identities, inboxes, thread search, and file leases.