docs(specs): sync and reorganize tasks.md for 004 and 005

[Matobi98]

Summary

Two docs chores from #122, both covering specs/*/tasks.md files that had drifted from reality.

specs/005-transport-v2-migration/tasks.md

All 12 tasks were still [ ] even though the transport migration merged in #111 and was verified end-to-end. Marked all 12 done, and added a status note at the top: the plan is now a historical record, not a map of the current code. Verified against current code, it's been superseded further than the issue assumed — mostro-core is now 0.14.1 (past the 0.13.1 target), gift_wrap.rs no longer exists (wrap_mostro_message/unwrap_mostro_message now live in rust/src/nostr/transport.rs), and T011's "leave gift wrap on peer/dispute chat untouched" no longer holds, since that chat also moved off gift wrap entirely in #246/#254.

specs/004-mostro-p2p-client/tasks.md

Three changes:

• Added an "Open work at a glance" summary near the top. Almost the entire file (T001-T145) is done, so phases 18/19/20 aren't really historical work mixed into an active backlog, they're just appended sequentially at the end. The real open work is two small clusters of partial tasks: file attachments (T079, T080, T081) and dispute admin chat (T088, T090-T093), both cases where the Rust side is ready but nothing in Dart calls it yet.
• Corrected an existing transport-v2 note that had gone stale in the same way as the 005 file: it claimed peer chat still uses NIP-59 gift wrap, which is no longer true after Migrate peer-to-peer chat to the gift-wrap-free envelope (protects against the gift wrap apocalypse attack) #246/feat(disputes): move the dispute chat onto the kind-14 envelope #254.
• Added a pointer task (T148, Phase 21) for the mnemonic-restore flow: it has a contract in contracts/identity.md (import_from_mnemonic with recover=true) but no task ever tracked it, per @grunch's note on Restore sessions and trades from mnemonic (currently impossible after reinstall) #142. Scoped as a pointer to the epic (Restore sessions and trades from mnemonic (currently impossible after reinstall) #142 and its sub-issues), not a new task breakdown, so it doesn't duplicate work already planned there.

Docs-only change, no code touched.

Closes #122

Test plan

• Read both files fully before editing to confirm the actual state (checked boxes, phase structure, existing notes) rather than assuming the issue's description was still accurate
• Cross-checked the corrected claims against the current code (mostro-core version, file layout, chat transport) the same way as in README: explicitly list supported NIPs and BUDs #123
• N/A, no Rust/Dart code changed, so cargo fmt/clippy/test and flutter analyze/test do not apply

[coderabbitai]

Important

• 🔍 Trigger review

This repository does not receive automatic reviews because it has fewer than 10 stars.

⚙️ Run configuration

Configuration used: Organization UI

Review profile: CHILL

Plan: Team

Run ID: 6ace3872-511e-4f34-96c1-8fb366fdc0b2

---

Thanks for using CodeRabbit! It's free for OSS, and your support helps us grow. If you like it, consider giving us a shout-out.

❤️ Share

• X
• Mastodon
• Reddit
• LinkedIn

_{Comment @coderabbitai help to get the list of available commands.}

[ermeme]

Reviewed current head 978d76d7e85b4a76d2de7ba88118389dc28bdef8.

Blocking documentation issues

This PR claims to synchronize the task plans with reality, but the new “Open work at a glance” omits a major existing gap: T010 remains marked [x] even though rust/src/db/indexeddb.rs still returns IndexedDB not yet implemented for orders, trades, relays, identity, and queued messages. The task text says the IndexedDB backend must support all 11 entities, so marking it complete and claiming that almost everything through T145 is done is materially inaccurate. Either mark T010 as partial/not done and include it in the open-work summary, or update the task wording/status to match the actual supported scope.

Also, the new historical transport note correctly says peer/dispute chat moved off gift wrap, but the corresponding historical T011 still says it was completed by leaving gift wrap untouched. If this file is intended as a truthful current-status record, annotate that task as superseded (as done in the 005 file) rather than leaving a completed task whose acceptance text is now false.

These are correctness blockers for a PR whose stated purpose is syncing the task documents. No CI checks are configured for this docs-only PR.

[Catrya]

@Matobi98 please fix the conflicts and address the hermemes review, thanks!

[Matobi98]

Thanks for the review — both blockers addressed in the rebased head.

1. Conflicts: rebased onto main (179 commits). The only clash was the Transport v2 note, which upstream had already rewritten in 61fa304 — I dropped my edit of that line and kept upstream's.

2. T010: now [~] with the real scope spelled out — only messages, settings and the active Mostro pubkey are backed by real object stores; orders, trades, relays, identity, queued messages and trade keys still return IndexedDB not yet implemented (Web: IndexedDB storage backend is a stub — nothing persists across a reload #233). Added as the first bullet in "Open work at a glance" (phase order), and fixed the surrounding claims: "Almost everything below (T001–T145) is [x]" → "Most of what follows…", and "two clusters" → "three clusters".

3. Superseded tasks: added (Superseded — see the Transport v2 note above.) inline on 004's T011 and T012, the two that name Kind 1059 explicitly; the rest of the historical set stays covered by the block note. 005's T011 already carried that annotation.