v3.56.10 — TUI pictures and looks, MCP project cwd and daemon fixes, memory and analyze fixes

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.