v3.56.10

Patch release: `@claude-flow/cli`, `claude-flow`, `ruflo` 3.56.10 (latest, alpha, v3alpha). The CLI pins `@claude-flow/memory` 3.0.5 (#4183) and `@claude-flow/cli-core` 3.7.3 (#3987), both published earlier, plus `shared` 3.0.5, `security` 3.0.5 and `codex` 3.0.5 exactly. Signed helpers changed in #4178 and were re-signed there.

### Added
- **`ruflo tui`** draws the console's pictures, looks and motion (console 0.44.x, #4182, #4188): `--look bbs|orbital|plain`, `--fps`, `--anim`, `--pictures`, `--boot`, and any console view with `--view`. It draws at the default size when the terminal reports 0 columns (#4160, @martinvlad).

### Changed
- `ruflo init`: the v3 implementation skills are opt-in (`--full` adds them), like `agents.v3` (#4140, #4139, @z4forzubair).

### Fixed
- **MCP server and daemons**: the server starts in the session project and ignores a `CLAUDE_FLOW_CWD` that names the home directory or a filesystem root (`cli-core` 3.7.3); user-level claims resolve through `os.homedir()` on Windows; duplicate or zombie daemons per workspace are stopped (#3987, @socrates6969). Trajectory counts persist across processes and daemon jobs feed learning (#3988, @socrates6969). One node process per MCP session, with `node:http` and the security barrel kept off the startup path (#3989, @socrates6969).
- Console reads the project on Windows (a drive path is absolute) (#4177, #4175). Neural data dir resolves through the project cwd; stored worker error text is capped.
- **Hooks and core**: failed tool outcomes are recorded, hooks run from the project root, telemetry stays off the tool path (#4178).
- **Memory**: calibrated default search threshold with near-miss hints (#4184, #4129); `memory_export` reads the CLI store as well as AgentDB (#4186, #4148); the AgentDB directory is created before the bridge opens it (#4155, #4132, @martinvlad); `TieredMemoryStore` recalls multi-word and natural-language queries (`memory` 3.0.5, #4183, #4152).
- **Intelligence**: long memories recalled for short route prompts (#4172, @rudycelekli); SONA reinforces the nearest pattern for near-identical tasks (#4185, #4134); AIDefence learns missed threats and autopilot learning loads (#4187, #4147, #4146).
- **Workflow and agents**: invalid gates and unsupported steps fail closed (#4164); agent health derives from observed execution outcomes (#4167) (@rudycelekli).
- **Analyze**: classification from diff content and commit subjects, safe Git revision ranges, encryption paths flagged security-sensitive, accurate reviewer-role docs (#4165, #4162, #4169, #4163, @rudycelekli).
- **Security scan**: interpolated `exec` calls after imports are detected (#4166, @rudycelekli); a failed `npm audit` reports "could not scan" instead of a clean tree (#4154, #4124, @martinvlad).
- Clear "not installed" message for `@ruvector/rvagent-wasm`; only new patterns are counted (#4179, #3979). Skills, plugin contracts and ruvector/neural/kg surfaces (#4181).

### Notes
- `@claude-flow/browser` (agent-browser bump, #4130) is a standalone package outside this train and was not republished.

### Thanks
@rudycelekli, @martinvlad, @proffesor-for-testing, @ronmikailov, @z4forzubair, @socrates6969, @nicholas-ruest and @pacphi for the fixes and reports across 3.56.6 to 3.56.10.