ruvnet / rvf-manage

RVF Management

Manage RVF files for portable, transferable agent memory.

When to use

When you need to export agent memory to RVF format for backup, transfer between projects, or share knowledge between teams.

Steps

1. List memories — call mcp__plugin_ruflo-core_ruflo__memory_list to see all stored memories
2. Export — npx @claude-flow/cli@latest memory export -o <file> [-n <namespace>] [-f json|csv|binary] (vectors included by default; --include-vectors false to drop them). Pick the output path inside the project, never an existing file you have not read
3. Import — npx @claude-flow/cli@latest memory import -i <file> [-n <namespace>] (merges and skips duplicates by default), or call mcp__plugin_ruflo-core_ruflo__memory_import_claude to pull in Claude Code auto-memories
4. Migrate — call mcp__plugin_ruflo-core_ruflo__memory_migrate for format upgrades
5. Stats — call mcp__plugin_ruflo-core_ruflo__memory_stats for storage metrics

RVF format

RVF (Ruflo Vector Format) stores:

• Vector embeddings (384-dim ONNX)
• Metadata (timestamps, namespaces, tags)
• Causal relationships between entries
• Session context and agent scope

Transfer between projects

Copy patterns from another local project (nothing leaves the machine):

npx @claude-flow/cli@latest hooks transfer from-project --source /path/to/other/project

hooks transfer store is different: it is the pattern marketplace on the public IPFS registry (list, search, download, publish). Never run publish for project memory unless the user asks for it, and scan the content for PII and secrets first (mcp__plugin_ruflo-core_ruflo__transfer_detect-pii). Anything published may be cached and cannot be recalled.

Exported files are not encrypted by the at-rest setting (see the plugin README); treat them as sensitive if the memory is.