…classifier

The #6211 token/context-limit banner keys off the chat surface error string,
which on current main is the output of the kernel's `classify_streaming_error`
(#6215 merged).
Two cases were wrong against that output:

- False negative: the canonical context-overflow message is collapsed to
  "Context is full. Try /compact or /new.", which matched none of the phrase
  list, so the banner missed the exact case it exists for.
  Add "context is full".
- False positive with harmful advice: an internal usage/spending-budget cap is
  reported as "Usage budget reached ... not a full context window — /compact
  will NOT help ...".
  That string contains "context window" (inside a negation) and so fired the
  banner, telling the user to start a new session — which does nothing for a
  persisted budget window.
  Suppress the banner for usage-budget wording.

Add regression cases pinning both behaviours (they fail without the fix) and
refresh the CHANGELOG entry to reflect that #6215 has merged and only adds a
usage indicator, not a per-turn error classification.

…mit (#6211) (#6231)

* feat(dashboard/chat): guide user to a new session on token/context limit

When the latest agent-chat turn fails with a token / context-window or
length / quota limit, the chat view now renders an inline guidance banner
with a one-click "Start a new session" action that reuses the existing
useCreateAgentSession mutation, instead of leaving only a raw error bubble.

Detection is a frontend heuristic over the daemon / provider error string
(isContextLimitError), because main carries no structured context-exhaustion
signal on the chat surface.
A structured signal can replace it later (related: in-flight #6215).

Scoped to the agent session chat view — channels are config-only surfaces in
the dashboard with no conversation UI.

Adds en/zh/uk locale strings and ChatPage.limit.test.ts.

Closes #6211

* refactor(dashboard/chat): trim multi-line comment blocks and remove issue refs from source

* fix(dashboard/chat): align context-limit heuristic with kernel error classifier

The #6211 token/context-limit banner keys off the chat surface error string,
which on current main is the output of the kernel's `classify_streaming_error`
(#6215 merged).
Two cases were wrong against that output:

- False negative: the canonical context-overflow message is collapsed to
  "Context is full. Try /compact or /new.", which matched none of the phrase
  list, so the banner missed the exact case it exists for.
  Add "context is full".
- False positive with harmful advice: an internal usage/spending-budget cap is
  reported as "Usage budget reached ... not a full context window — /compact
  will NOT help ...".
  That string contains "context window" (inside a negation) and so fired the
  banner, telling the user to start a new session — which does nothing for a
  persisted budget window.
  Suppress the banner for usage-budget wording.

Add regression cases pinning both behaviours (they fail without the fix) and
refresh the CHANGELOG entry to reflect that #6215 has merged and only adds a
usage indicator, not a per-turn error classification.

---------

Co-authored-by: Evan <[email protected]>
Co-authored-by: Claude <[email protected]>

Merge upstream/main into the BossFang fork; origin/main was 15 commits behind.

Notable upstream changes:
- librefang#6225 scope the compaction-summary banner to the compacted session (adds canonical_sessions.compacted_summary_session_id).
- librefang#6196 deny WASM fs_write to the audit anchor via a capability deny-list.
- librefang#6226 / librefang#6227 add agent label and exit-reason metrics to the agent loop.
- librefang#6215 context-window usage indicator + honest quota-error classification.
- librefang#6217 per-instance sidecar secrets so each agent owns its own handle.
- librefang#6194 global Auto-Dream on/off switch on the Memory tab.
- librefang#6211 / librefang#6214 / librefang#6212 token/context cap fixes; librefang#6208 refuse to delete the active prompt version.
- librefang#6218 browser_tools.rs ToolError migration; librefang#6203 launchctl let-binding build fix; librefang#6224 docs dep bumps; librefang#6193 drop five orphaned email deps.

Conflict resolution:
- deny.toml: keep our RUSTSEC-2025-0141 (bincode 2.0.x) ignore entry (take ours).
- deny.toml: ignore three pre-existing unmaintained advisories newly tripped by the RustSec DB (all transitive via the UAR provider chain, none introduced by this merge): RUSTSEC-2024-0384 (instant), RUSTSEC-2024-0436 (paste), RUSTSEC-2025-0119 (number_prefix).

BossFang preservation (SurrealDB schema parity for upstream SQLite v46 / librefang#6225):
- Add crates/librefang-storage/src/migrations/sql/032_canonical_sessions_compacted_summary_session_id.surql declaring the new field on the SCHEMAFULL canonical_sessions table (SCHEMAFULL silently drops undefined fields on write), registered as version 32 in migrations/mod.rs.
- migrate/sqlite_to_surreal.rs: copy the new column so a SQLite to SurrealDB migration preserves the owning-session pointer rather than dropping it.
- backends/surreal_session.rs: preserve compacted_summary_session_id across canonical appends, since the upsert replaces the whole record.
- Cargo.lock: pick up the workspace version bump (beta.19 to beta.20) the merge introduced.

Verification:
- cargo check --workspace --lib — clean.
- cargo check -p librefang-storage -p librefang-memory -p librefang-uar-spec — clean.
- cargo test -p librefang-storage migration — ok (migration ordering / SurrealDB-3 flexible-syntax invariants).
- cargo test -p librefang-memory --lib session — 50 passed (incl. upstream's store_llm_summary round-trip test).
- cargo clippy -p librefang-storage -p librefang-memory — clean.
- python3 scripts/enforce-branding.py --check — clean; Tauri desktop audit and URL-drift scan both clean.