Ship the batch: undo, parallel subagents, lazy MCP, hot-reloaded skills; docs and CI/CD

Loop and ergonomics batch across Now/Next and Maintenance:
- /undo and /redo via pre-prompt file snapshots (snapshot.ts)
- task takes a tasks[] array and runs investigations concurrently (subagent.ts)
- lazy MCP tools: mcp_list/mcp_inspect/mcp_call meta-tools, eager opt-in (mcp.ts, config.ts)
- skill tool reads its list live so a mid-session install is callable next turn (skills.ts)
- tool-name lists (tool-kinds.ts) derived from a mutating() marker; gates previously ungated writes
- prune/session recovery path summarized, and step-back doom-loop primitive (step-back.ts)
- @file completion re-walks on a slow cooldown; estimateTokens and pricing labeled as estimates

Docs: README, CHANGELOG, docs/{mcp,architecture,development} updated to match.
CI/CD: bun install-store cache and concurrency gates on both workflows; release.yml now
composes file-based release notes via scripts/make-release-notes.ts and verifies every binary.

src/pricing.ts

@@ -4,6 +4,10 @@ export type Rate = { inputPerMTok: number; outputPerMTok: number };
  * USD per million tokens. Prefix match on the model id, longest first, so
  * `claude-sonnet-4-5-20250929` resolves via `claude-sonnet-4-5`. Published rates
  * drift, so this is a best-effort estimate rather than a billing source.
+ *
+ * Source: vendor pricing pages, checked 2026-09-17. Anthropic (Anthropic API, not
+ * Batch) and OpenAI listed rates; DeepSeek and Grok per their API pricing. Rates
+ * are for input, then output. Re-verify before trusting a live spend figure.
  */
 const RATES: Record<string, Rate> = {
   'claude-opus-4': { inputPerMTok: 15, outputPerMTok: 75 },