Dotfile Sync Manager

ReviewAudited by ClawScan on May 10, 2026.

Overview

This instruction-only sync skill appears purpose-aligned, but it can change dotfile/MCP/Syncthing configurations and write persistent project memory, so review actions before running them.

Install or use this only if you want an agent to help manage dotfiles, MCP server configs, Syncthing settings, and project memory. Before allowing actions, review chezmoi diffs, confirm Syncthing paths and devices, verify MCP servers are trusted, and inspect any knowledge before it is saved to Serena memory.

Findings (6)

Artifact-based informational review of SKILL.md, metadata, install specs, static scan signals, and capability signals. ClawScan does not execute the skill or run runtime probes.

What this means

If run with the wrong path or settings, the agent could alter what Syncthing shares across devices.

Why it was flagged

The skill documents direct Syncthing REST API calls that can add or change a synced folder. This is expected for Syncthing setup, but it is still a powerful configuration action.

Skill content
curl -X POST -H "X-API-Key: $API_KEY" ... http://localhost:8384/rest/config/folders ... "path": "~/.local/share/chezmoi"
Recommendation

Review the folder path, folder ID, and Syncthing settings before allowing the command to run.

What this means

Running these commands can disrupt Syncthing temporarily and remove the local index backup after verification.

Why it was flagged

The DB reset procedure stops a service, moves the Syncthing index, and later deletes the backup. The guide scopes this to Syncthing and says to remove the backup after verification, so it is a note rather than a concern.

Skill content
brew services stop syncthing ... mv ~/Library/Application\ Support/Syncthing/index-v2 ... index-v2.bak ... rm -rf ~/Library/Application\ Support/Syncthing/index-v2.bak
Recommendation

Only use the reset workflow for a real stale-index problem, verify sync health first, and keep the backup until recovery is confirmed.

What this means

Anyone controlling these commands could query or modify the local Syncthing instance using its API key.

Why it was flagged

The skill reads the local Syncthing API key and uses it for localhost REST calls. This is purpose-aligned, but the key grants privileged access to Syncthing configuration.

Skill content
API_KEY=$(xmllint --xpath '//configuration/gui/apikey/text()' ~/Library/Application\ Support/Syncthing/config.xml)
Recommendation

Use this only with trusted agents, avoid exposing command logs containing the key, and confirm requests go only to the intended localhost Syncthing endpoint.

What this means

Sensitive project details, outdated assumptions, or mistaken conclusions from a session could be saved and influence future agent behavior.

Why it was flagged

The skill can extract session knowledge and write it into Serena persistent memory, including overwriting existing memory. This is the stated purpose, but it creates durable context that future tasks may reuse.

Skill content
Automatically extracts knowledge from a specific session using MCP ... Use mcp__serena__write_memory ... Overwrite if existing memory exists
Recommendation

Review extracted knowledge before saving, avoid storing secrets, and be careful when overwriting existing memory.

What this means

Project/session content may pass through registered MCP tools before being stored.

Why it was flagged

The workflow bridges MCP/UTCP/code-mode tools to move knowledge between a sessions service and Serena memory. This is disclosed, but trust and permissions are controlled by those external MCP registrations.

Skill content
Call mcp__code-mode__call_tool_chain tool: tool_call_chain: "claude_sessions.claude_sessions_mcp_extract_project_knowledge" ... "serena.serena_write_memory"
Recommendation

Confirm the registered MCP templates and servers are trusted before syncing knowledge through them.

What this means

A bad MCP server entry or unintended config change could break or alter tool behavior across several apps.

Why it was flagged

A single MCP server list change is designed to propagate into several agent/tool configurations. This is coherent with the sync purpose, but mistakes can affect multiple applications at once.

Skill content
After editing, `chezmoi apply` → Automatically applied to all apps: `~/.claude.json`, `~/.cursor/mcp.json`, `~/.gemini/antigravity/mcp_config.json`, `~/.utcp_config.json`
Recommendation

Inspect the chezmoi diff before applying, validate new MCP servers, and keep a rollback path for generated configs.