Epic: TinyClaw OpenClaw parity via Terraphim extensions

Start with docs/plans/tinyclaw-openclaw-parity-research-2026-02-27.md and docs/plans/tinyclaw-openclaw-parity-design-2026-02-27.md, then review the linked tracks #588, #589, #592, #593, and #591. Done means the listed parity tracks are implemented through existing Terraphim functionality, with practical core-channel workflows, consistent safety controls, and documentation matching shipped behavior.

Summary

Track terraphim_tinyclaw capability parity with OpenClaw by extending Terraphim crates and shared patterns, not duplicating OpenClaw internals.

Principle (Non-Negotiable)

• Extend existing Terraphim functionality first.
• Do not introduce parallel/duplicate subsystems when a Terraphim crate or module already solves the problem.

Reference Docs

• Research: docs/plans/tinyclaw-openclaw-parity-research-2026-02-27.md
• Design: docs/plans/tinyclaw-openclaw-parity-design-2026-02-27.md

Scope (Top 5 Tracks)

• #588 Foundation hardening (slash/session correctness, config wiring, unified guardrails)
• #589 Provider-backed web search/fetch quality
• #592 Markdown-defined commands and skills via shared Terraphim parsing/runtime
• #593 Voice transcription pipeline with graceful fallback
• #591 Orchestration extensions (session tools + spawn integration + cron sequencing)

Existing Related Issues (extend, don’t duplicate)

• #560 TinyClaw: Add agent spawning tool via terraphim_spawner
• #561 TinyClaw: Document standalone architecture

Exit Criteria

• TinyClaw users can achieve practical OpenClaw-like workflows on core channels.
• Markdown-defined commands/skills are first-class.
• Safety controls are consistent across tools and skills.
• Documentation matches shipped behavior.