compress: route every trigger path through compress_and_save_session ...

- CompressionWorker keeps the real-token gate (ctx.context_tokens) but now
  delegates to compress_and_save_session(reason="postturn") — it used to
  call compress_context directly with its own save logic and silently
  no-op'ed on any summarizer failure, leaving the session over the
  threshold until the next turn's gates
- pre-turn and /compact gates use real_baseline_estimate instead of the
  chars//3 heuristic (mid-turn already did), so code-heavy contexts
  compress in time instead of tripping ContextTooLargeError
- summarizer input keeps head + tail instead of head-only truncation,
  which silently dropped the newest summarized messages closest to the
  keep window
- critical tool results over the 4000-char budget keep head+tail halves
  instead of collapsing to the 800-char non-critical preview (a 3999-char
  read survived whole, a 4001-char one lost 98%)
- images count 500 tokens each, not 500 per message carrying them
- _plan_compression is the single partition decision shared by
  compress_context and would_compress, so the dry-run prediction cannot
  drift from the real path; the trigger reason is now logged
  (preturn/midturn/postturn/forced)

profiles: retire gate flags and retrain prompts for the plan tool ...

- drop planning_enabled / planning_mandatory / planning_phase2_enabled /
  observe_skips_plan_enabled / adaptive_replan_enabled from AgentProfile,
  the loader and admin serialization
- add `plan` to native tools of developer, navi_code, tool_developer,
  secretary, server_admin, modeler_3d — without the gate they would
  otherwise lose planning entirely
- system prompts: call plan for non-trivial multi-step work before
  execution; complex plans get presented for confirmation; stuck or goal
  changed -> plan with a reason

planning: replace the mandatory pre-turn gate with an agent-invoked plan tool ...

- Phase 2 (critique) retired; observe/MODE and the top-level DIRECT shortcut
  removed (sub-agents keep DIRECT for trivial subtasks)
- replan merged into plan: `reason` (+ updated_goal) packs the re-plan
  context, without a reason it is fresh planning
- both phases call the LLM with think=False — cloud reasoning models leak
  their chain-of-thought into structured output on non-streaming calls —
  and strip the gemma "thought<channel|>" content artifact
- Phase 1 conversation windowed to ~20k chars, newest-first with the
  original task pinned (28.7k-token prompts came back as 1-token output)
- confirmation by COMPLEXITY: the tool result tells the agent to present
  complex plans and wait; sub-agents keep the execute-now injection
- PlanningStatus/PlanReady reach the UI mid-turn via current_event_sink
  (ctx.event_sink is None in the agent loop — events silently never
  reached the WS before)
- agent gate, casual-message detector and adaptive re-plan nudges removed;
  anti_stall keeps only stall detection

Merge branch 'master' into feature/navi-code ...

# Conflicts:
#	navi/api/websocket.py
#	navi/config.py
#	navi/core/agent.py
#	navi/core/orchestrator.py
#	navi/main.py

navi_code: add gemma4 31b/26b qat models to the profile fallback chain

security: critical batch 1 — RCE/XSS/CORS/webhook hardening ...

Backend:
- auth/deps: fix refresh-lock clock race (cleanup now monotonic like the cache)
- core/registry: add missing structlog logger (NameError on fallback path)
- config: GNAUTH_WEBHOOK_SECRET, NAVI_ALLOWED_ORIGINS (+list property)
- webhooks: HMAC-SHA256 signature verification (503 unconfigured in auth mode,
  unsigned+warning in no-auth mode, 403 bad/stale signature, 400 bad JSON)
- main: CORS from NAVI_ALLOWED_ORIGINS in auth mode (fail-fast on empty),
  eval router behind require_admin, /debug only registered in no-auth mode
- auth routes: mobile-done sid validation (32 hex) + CSP header + safe JS
  escaping (reflected XSS); Secure cookie flag via helper (https base URL)
- websocket: anonymous WS rejected in auth mode (closes legacy-session RCE);
  socket registered only after access checks; stop_session auth gate
- messages: REST agent start now takes session lock + busy flag (409 on
  active run), contextvars reset in finally

Webclient:
- useMarkdown: DOMPurify.sanitize on all rendered markdown (stored XSS),
  image-URL scheme whitelist, delegated error listener (no inline onerror)
- html.html artifact viewer: sandbox without allow-same-origin (opaque origin)
- tests: DOMPurify runs under jsdom (happy-dom Node.prototype.nodeName getter
  breaks DOMPurify); useWebSocket tests get localStorage stub + ui-kit alias

Tests: pytest 1049 passed, 1 skipped; vitest 61 passed

navi_code: add web lookup, ssh, and image_view — grow beyond local-only ...

navi_code was strictly local: no web, no remote hosts, no image viewing. That
made it lean on the developer profile whenever a coding task needed a doc
lookup or a remote command. Bring those capabilities in so navi_code covers
full coding work itself, while staying terminal-first and local-by-default.

config: agent gains image_view, ssh_exec, and the navi-web MCP (search/browse/
request); subagent gains image_view + navi-web (no ssh_exec — sub-agents don't
reach remote hosts, matching developer). Description/short_description/
full_description updated to "terminal-first with web and remote access". Also
includes the gemma4:12b-it-qat-128k model in the model list (local 128k-context
option alongside 31b-cloud).

prompt: soften "Local-only, no remote hosts" to "local by default, with web
lookup (docs/APIs) and ssh (remote ops) when the task needs it". Add compact
Web lookup and Remote access sections under Execution environment, and update
the sub-agent toolset briefing to list image_view + web and to exclude ssh_exec.
scope_boundary still applies; ssh destructive/system-wide actions need user
confirmation. Kept additions minimal to avoid bloating the prompt for the 12b
model.

navi_code: stop delegating code work to the developer profile ...

navi_code was systematically spawning sub-agents as `developer` instead of
itself, even though it already covers local code work. Root cause was a
three-channel nudge: the navi_code prompt recommended `developer` for "general
code work" and framed omit (-> navi_code) as the exception; spawn_agent's
description listed `developer` as the coding example and never mentioned
navi_code; the injected Available-profiles list shows developer's broad
"General-purpose software development" blurb. The planner locks profile_id at
plan time on those same signals, so the bias is baked in before execution.

Flip the default in all three places: for code work OMIT profile_id so the
sub-agent runs as the current profile (navi_code); set profile_id only for a
different specialisation (secretary/server_admin/tool_developer). developer is
not removed as a profile — it is just no longer recommended from navi_code.

code_exec: agent-controlled timeout (30s default, 300s max) + language metadata ...

The 30s timeout was a hard constant; long computations or test suites hit it
with no escape hatch. Expose it as a `timeout` param (clamped 1-300s, default
30) like terminal's. Also stamp `language: "python"` into metadata so the
TUI renderer can highlight without hard-coding, and record `timeout` on the
timeout result for a dedicated timeout status card.

compress: raise budget 6000 + richer prompt detail + input headroom ...

Increase the compression summary budget and push the prompt toward more
detail and completeness, so the agent keeps more durable facts when old turns
are folded into a summary.

Budget (output):
- context_summary_max_tokens 4000 → 6000 (global default).
- navi_code compression_max_tokens 4000 → 6000 (profile override).

Prompt (base, all profiles):
- New 'Intermediate Findings' section: durable facts from read/grep/log/terminal
  that informed decisions (was forbidden by 'do not preserve intermediate
  reasoning'). Existing sections get richer instructions (Active Files: key
  symbol affected; Decisions: rationale + trigger; Completed: step + file +
  verification outcome; Errors: snippet + what was tried + outcome).
- Loosen 'tight bullet points' → 'specific and complete; prefer listing over
  generalizing — losing a fact costs more than a longer summary'.
- 'Do not write implementation code' → 'preserve short critical code verbatim
  (final signatures, one-line fixes, key config lines); drop long blocks'.

Input headroom (so the summarizer can SEE what it must preserve):
- Non-critical tool result preview 300 → 800 chars.
- _MAX_SUMMARY_INPUT_CHARS 24000 → 32000 (fits 65k–128k windows with 6k output).

Profile prompt (navi_code/compression_prompt.txt): preserve short diagnostic
snippets verbatim (exact error line, failing assertion, one-line fix) instead
of blanket 'do not preserve long output'; keep code signatures + short key
snippets, not just names.

Tests: Intermediate Findings section in base prompt; 'short critical code'
phrasing; non-critical preview keeps 800 chars; meta-summary threshold raised
(2×6000 chars). Fixed flaky test_attach_session_seeds_context_fill (relied on
_startup worker timing after Etap 4 lengthened attach; now explicit attach
like its sibling + mock get_todos to skip a 30s×2 real-network timeout).
Full suite: 1008 passed, 1 skipped.

todo: add op — append steps discovered mid-task (preserve statuses) ...

The todo tool only had set/view/update/clear. Adding a step that surfaced
mid-task meant either 'set' (which replaces the plan and resets every status
to pending) or 'replan' (1–3 LLM calls) — both heavy for a single new step, so
the agent usually skipped recording it. Add a cheap 'add' op:

- navi/tools/todo.py: op 'add' (tasks: list[str]) appends new _Task entries to
  the end of the existing plan in pending status; existing steps and their
  statuses are untouched. Requires an existing plan (points to 'set'
  otherwise). Schema enum + op description + tool description updated.
- system_prompt.txt: tell the agent to record new subtasks with 'todo add'
  right away (not held in head, not a replan); clarify the 'small todo edit'
  line — add via 'todo add', drop/merge/reorder via 'set' (re-apply statuses).
- renderers/todo.py: TodoStartedRenderer renders the add card
  ('→ todo · add (N)') listing the new pending steps, not a JSON dump.

Tests: add appends + preserves statuses / requires existing plan / requires
tasks; renderer add card + empty-tasks. Full suite: 1005 passed, 1 skipped.

terminal: REST list/close endpoints + async client helpers (Etap 2) ...

The /terminals modal and the status-bar count need a server source for the
session's persistent terminals (background dev servers, etc.) and a way to
force-close one. Add REST endpoints over the existing TerminalManager:

- GET /sessions/{id}/terminals → terminal_manager.list(session_id) →
  {session_id, terminals: [summary,...]}. Access-checked (navi.sessions.read_all).
- POST /sessions/{id}/terminals/{name}/close → terminal_manager.close(...)
  → {session_id, terminal_name, closed: bool}. Name is URL-encoded by the
  client (terminal names may contain spaces).
- clients/terminal/api.py (async): list_terminals(session_id),
  close_terminal(session_id, name) — for the TUI modal + status-bar seed.

Tests: list empty / list returns active / close routes to manager / 404 on
unknown session (fake TerminalManager injected into the container, mirroring the
kv_store injection pattern). Full suite: 972 passed, 1 skipped.

Plan: docs/terminal_tool_plan.md — Etap 2 of 5.

terminal: session-broadcast sink + lifecycle events (Etap 1) ...

Persistent terminals emitted their stream only into a per-tool-call sink, which
closed once the open action returned — so a background dev server's output
stopped reaching the client immediately after open. TerminalClosed was never
emitted at all (the handler existed, the emitter did not).

Replace the per-tool event_sink with a single callback the orchestrator wires to
the session's WebSocket(s):

- TerminalManager: set_event_callback(cb) + _emit(session_id, event). Reader
  tasks and open/close deliver TerminalOutputDelta / TerminalOpened /
  TerminalClosed through the callback, which lives for the terminal's (and
  session's) lifetime — not the tool call's.
- New TerminalOpened event (terminal_opened: name/description/pid/background).
  TerminalClosed now actually emitted from _close_one (explicit/idle/shutdown/
  session_ended). Background output keeps streaming after open returns.
- terminal.py: drop the per-tool event_sink from _tm.open (no longer needed).
- orchestrator._on_terminal_event → _notify_session (per-session WS fan-out,
  same path recall/mcp updates use). container wires the callback after building
  the orchestrator.
- Events imported lazily inside terminal_manager methods to avoid a circular
  import (navi.core.events → navi.core.__init__ → agent → registry → navi.tools).

Tests: open emits TerminalOpened; background output streams via the callback
after open returns (gap-3 fix); close emits TerminalClosed; no callback does not
break readers. Existing terminal tests unchanged. Full suite: 968 passed, 1 skipped.

Plan: docs/terminal_tool_plan.md (5 etaps, this is Etap 1).

mcp: dedicated runner task per client (fix cross-task cancel-scope on shutdown) ...

MCP client SDK transports (stdio/sse/streamable_http) + ClientSession are
anyio task groups whose cancel scopes require __aenter__/__aexit__ in the
SAME asyncio task. McpClient entered the transport in one task (lifespan
connect / health-check reconnect / request retry) and exited it in another
(lifespan teardown) -> RuntimeError: Attempted to exit cancel scope in a
different task than it was entered in.

Refactor McpClient to own a single long-running runner task that holds the
AsyncExitStack and performs ALL transport enter/exit + list_tools/call_tool.
The public async API (connect/disconnect/list_tools/call_tool/mark_disconnected)
just enqueues a _Cmd and awaits a Future, so callers from any task no longer
cross cancel-scope boundaries. connected/instructions mirror from the runner
onto the instance to stay sync-readable. disconnect() enqueues a stop command
and awaits shield(runner) so teardown isn't interrupted by lifespan cancel.

Also call mcp_manager.stop_health_check() BEFORE disconnect_all() in
AppContainer.shutdown() so the health-check task cannot enqueue onto a
client whose runner is being torn down. mark_disconnected() is now async
(queued) and its manager caller updated.

Regression test: connect in one task, list_tools in a second, disconnect in
a third — the exact scenario that raised the RuntimeError before.

compressor: target hysteresis — shrink to 65% after compression, not just below the trigger ...

Compression triggered at 90% but left the context just under the trigger, so a
fixed keep_recent (navi_code: 12 turns ≈ 104 messages) re-triggered a couple
of messages later — the context yo-yoed at the trigger line instead of gaining
headroom. Add context_compression_target (0.65): in turn-based (preturn) mode
compress_context shrinks keep_recent until the verbatim kept region fits 65% of
the window, folding the extra turns into the same single summary LLM call; a
safety net in compress_session token-budget-truncates if the kept region alone
still exceeds the target (midturn kept a huge in-flight turn, or the
keep_recent floor can't fit). Trigger 90% → target 65% leaves real headroom.

filesystem: strip trailing newline in _number_diff so diff lines aren't blank-separated ...

_unified_diff (used by edit_lines / smart_edit) feeds difflib with
"line + \n" so difflib does not emit a "\ No newline at end of file" marker.
With lineterm="" difflib keeps that trailing "\n" on content lines, so the
final "\n".join produced "\n\n" between every diff line — a blank line
between each, bloating the model-facing output and rendering with double
spacing in both clients. Strip the trailing "\n" in _number_diff before
processing; no-op for the _diff action, whose lines come from splitlines()
without trailing newlines.

Автономность мелких моделей: OUTPUT DISCIPLINE, milestone-todo, перехват финала хода ...

Два последовательных захода над одной проблемой — мелкие модели (12–30B) в navi_code
теряют автономность на трудных шагах: «остановился поболтать» вместо действия и
слишком большое расстояние между пунктами плана.

Заход 1 (З1–З3):
- З1: OUTPUT DISCIPLINE в системном промпте navi_code — «act, don't announce»,
  без few-shot антипримеров. Контракт хода: ответ без tool_calls = конец хода, поэтому
  объявление намерения текстом убивает автономность; правила заставляют вызывать
  инструмент в том же ходе.
- З2: плоский todo + метка группы milestone + декомпозиция. _parse_plan_steps
  возвращает list[tuple[milestone, text]]; milestone — метка группировки (не сущность,
  без статуса), «done» вычисляется при рендеринге; подшаги = больше плоских шагов
  (без вложенности). TUI side-panel группирует по milestone (плоский фолбэк при пустом
  milestone). Plan depth: max 15→20 + правило декомпозиции.
- З3: adaptive re-plan «длинный шаг» — nudge «разбей шаг» при in_progress ≥ порога
  итераций без смены todo (порог 4, раньше общего anti-stall warning на 8).

Заход 2 (шаги 1–3, после cloud-теста 31b vs 12b):
Корневая структурная причина: весь спасательный механизм (anti-stall warning с явным
предложением reflect, adaptive re-plan) живёт только внутри tool-цикла — nudge
инжектируется в pre_turn *следующей* итерации, которой при «остановился поболтать»
нет (ход закрылся по return до post_turn). 31b застревала через «продолжаю
tool-итерации» → дожала до warning → спаслась; 12b — через «замолчала текстом» → мимо
всех nudge.

- Шаг 1: перехват финала хода. Если модель выдала bare-text, но в todo есть открытые
  шаги (pending/in_progress) и лимит не исчерпан — НЕ эмитить StreamEnd, а сохранить
  ассистентский текст в session.context, поставить системный nudge и continue
  (без StreamEnd, без workers — консистентно с multi-iteration tool-турами). Счётчик
  final_interceptions на AgentTurnContext, лимит final_intercept_limit (default 2),
  эскалация жёсткости (мягкий → «second stop»). has_open_steps в todo.py: пустой
  todo → False (защита casual-сообщений), failed/skipped терминальны. Профильные
  флаги final_intercept_enabled/limit в base.py + loader + admin.
- Шаг 2: жёсткий reflect-триггер в промпте — «~3 tool attempts on the same step
  without progress → call reflect IN THIS TURN (tool call, not reasoning aloud)».
- Шаг 3: открыть replan для застревания — «call replan when reflect showed the whole
  approach is dead (not one failed step, but the approach itself)».

Тесты: 874 passed, 1 skipped. Новые — has_open_steps (5), final intercept (5),
milestone-группировка, adaptive long-step nudge, парсер шагов с milestone-маркером.

cloud: navi_code model → gemma4:31b-cloud для тестирования догадки (31b признала
застревание, 12b — нет); .env cloud-host уже gitignored.

Co-Authored-By: Claude <noreply@anthropic.com>

config: raise context compression threshold 0.70 -> 0.90 ...

Midturn compression now splits the in-flight turn correctly, so it's safe to
let the context grow closer to the limit before compressing. 0.90 gives the
model more raw context per turn and trips the compressor less often.

Worker test context_tokens bumped 50_000 -> 60_000 (76% -> 91.5%) to stay
above the new threshold; other compression tests pass the threshold as an
explicit argument and are unaffected.

Co-Authored-By: Claude <noreply@anthropic.com>

filesystem: show real file line numbers in unified diffs ...

_number_diff parses @@ hunk headers and prefixes each content line with its
real file line number: removed/context lines use the old (from) number,
added lines use the new (to) number, in `{marker} {num}│ {content}` form
(marker first so existing startswith highlighting stays valid; column
right-aligned to the largest line number). Hunk/file headers and the
`\ No newline` marker are left unchanged.

Applied in _unified_diff (covers edit / edit_lines / smart_edit) and the
`diff` action — a single server-side change, so both the TUI and the web
client receive numbered diffs in the tool result text.

TUI highlight_unified_diff renders the `{marker} {num}│` column dim and the
content in the marker color, so the number reads as meta, not as part of the
added/removed text. Lines without the prefix (standalone `diff` event, older
callers) are still colored whole.

Co-Authored-By: Claude <noreply@anthropic.com>

compressor: split in-flight turn in midturn mode + flush incrementally ...

Two related fixes for long autonomous turns in planning profiles (navi_code):

A1 — partition_messages turn-based branch held the in-flight (current) turn
verbatim when turns > keep_recent, so midturn compression was shallow
(e.g. 150 -> 141 messages). Now, when keep_recent_messages is set, the
in-flight turn is split like partition_current_turn_messages: head (user
request) + tail (recent tool steps) kept, middle summarized. The adaptive
swap is disabled in midturn mode (it could move the in-flight turn into
old_turns and summarize the current request whole). First branch
(len(turns) <= keep_recent) is untouched, so locked-in midturn tests hold.

B1 — agent.run_stream had no try/finally around the for-loop, so an
asyncio.CancelledError (server restart/shutdown) unwound the stack with no
flush: all in-memory turn messages (sequence_number < 0) were lost, only the
user message survived. Add incremental save() after planning, after the
assistant tool-call decision, and after each tool result, so a crash loses at
most the single in-flight tool call, not the whole turn. B2 (try/except
safety-net) was dropped: B1 leaves no window where an append is unsaved.

Tests: midturn split with many turns (partition + compress_session), and
crash/cancel persistence via a snapshot session store that mirrors the DB
boundary.

Co-Authored-By: Claude <noreply@anthropic.com>

permissions: remove broken client-side gate; add backend design doc ...

The terminal-TUI permission gate was a race: _execute_tools_with_sink
emits ToolStarted and starts the tool on the backend immediately, so by
the time the client dialog appeared the destructive action had already
run -- 'deny' could only stop the session, not un-execute the tool. It
also existed only in the terminal TUI (webclient/android had nothing),
and the system-prompt 'Strict Confirmation' nudge was an unreliable
duplicate.

Remove the broken machinery for a clean slate and plan the real
replacement from zero:

- delete clients/terminal/tui/permissions.py, screens/permission_dialog.py
  and their tests
- strip tui_app.py (engine, _deny_tool, _show_permission_dialog,
  _confirm_shell_command, _stop_session_worker, on_permission_request,
  tool_started gate) and the PermissionRequest TUI event; !cmd now runs
  directly (user-typed, no agent gate)
- drop the 'Strict Confirmation' prompt rule from navi_code/developer/
  tool_developer
- add docs/permissions.md: authoritative backend gate design (engine +
  registry + event + endpoint + agent gate + postgres policy + global
  rules config, sub-agents gated) -- design only, not yet implemented
- update docs/index.md, navi_code_cli.md, profiles.md, testing.md

Until Phase 1 lands there is no destructive-action confirmation -- a
conscious period without the false security of the old race-gate.

replan: integrated mid-task re-planning tool ...

Add a replan tool that re-runs the planner over the live session context
+ todo + scratchpad when the plan's structure is stale due to discoveries
(NOT failed steps -- that stays [Adaptive re-plan]). Integrated approach:
PlanningEngine.run gains is_replan/replan_context (suppresses DIRECT
shortcut and observe-skip, frames Phase 1 as a revision); ReplanRunner
packs reason/goal/todo/findings/errors and captures PlanReady; the tool
is exposed to navi_code/developer/tool_developer via a
current_replan_runner ContextVar set per-iteration in run_stream (correct
after switch_profile). New plan replaces the todo. Lazy events import
breaks the navi.tools -> navi.core -> navi.tools cycle.

profiles: port context-org/planning instructions to developer & tool_developer ...

Port the context-organization and planning-machinery framing accumulated
in navi_code to the developer and tool_developer profiles, adapted per
profile:

- Working state & memory (todo/scratchpad/context_transfer/schedule_recall/
  reflect/memory) replaces the stale "Context drift recovery" section.
- System signals subsection names the runtime-injected messages each profile
  actually receives: [Goal anchor], [Anti-stall warning], [Adaptive re-plan]
  (adaptive_replan is on for both), [Iteration N/M]. [Scope boundary] is
  omitted — scope_boundary_enabled is off on these profiles.
- Reading & searching, Editing policy (tool_developer), Safety Rules, Git
  discipline, and Project environments (isolated venv) added.
- developer: Workflow rewritten planner-aware with an observe carve-out
  (observe_skips_plan now on); Project knowledge replaced by docs-first
  Documentation; sub-agent briefing gains context_transfer + restricted-toolset
  bullets.
- tool_developer: keeps its MCP-specific 10-step workflow and prerequisites;
  sub-agent briefing gains context_transfer + restricted-toolset bullets
  (sub-agent cannot reload_tools/test_mcp_tool/mcp_status — run those inline).
- config: observe_skips_plan_enabled=true on both, so observe requests skip
  phase3 (no plan/todo for read/explain/inspect) — enables the Workflow
  observe carve-out and saves an LLM call on info requests.

Co-Authored-By: Claude <noreply@anthropic.com>

navi_code: prefer isolated project envs and bootstrap NAVI.md when absent ...

Add a "Project environments" section steering the agent toward the
project's existing isolated env (venv/uv/node_modules/target) — use it,
don't duplicate or bypass it — and only create a project-local one when
deps are needed and none exists; never install system-wide (system-wide
changes still require explicit confirmation). Also nudge the agent to
create NAVI.md when it's absent after real orientation work on a
non-trivial task, seeding the pointer structure so the next session
starts oriented.

Co-Authored-By: Claude <noreply@anthropic.com>

navi_code: enable planning phase 2 (structured review) ...

phase1 already emits REFLECT: yes|no, but with phase2 off the flag was a
dead-end. Enabling phase2 runs a Critic/Pragmatist/Detailer review pass
before phase3 whenever phase1 flags the task as complex (REFLECT: yes),
feeding "Plan Adjustments" into the execution plan. Skipped for simple
tasks and subagents (gated on `needs_reflect and not is_subagent`), so
the extra LLM call only lands on complex parent-agent work.

Resolves F8 machinery-gap G.

Co-Authored-By: Claude <noreply@anthropic.com>

navi_code: frame planning machinery in system prompt (F8) ...

Workflow "Plan" no longer duplicates the planner (phase1/3 already build a
structured plan and auto-populate todo); adds an observe/act carve-out so
observe requests don't get pushed to create a todo. New "System signals
you'll see" subsection names the runtime-injected messages ([Goal anchor],
[Scope boundary], [Anti-stall warning], [Iteration N/M]) so the model
recognises them as machinery and responds correctly — and corrects the
factual error that the goal anchor reads scratchpad `goal` (it reads the
original request + live todo). Closing paragraph warns that thinking isn't
re-injected and the plan's per-step executor assignments are lost to
compression, so conclusions/assignments should go in scratchpad.

Co-Authored-By: Claude <noreply@anthropic.com>

navi_code: give the sub-agent operational wisdom (editing/reading/todo) ...

By default (inherit_system_prompt=False) the sub-agent got only the 17-line
subagent_system_prompt — none of the parent's editing/reading/context
discipline. So a sub-agent with a clean context would still burn it: pulling
whole files into read, over-using smart_edit, not tracking steps. The clean
context is the sub-agent's main advantage — keep it clean.

Expand subagent_system_prompt.txt with a compact operational core:
- Editing: prefer edit/edit_lines (deterministic); smart_edit as last resort
  (extra LLM call, reads the whole file) — the same policy the parent now has.
- Reading: info before read, offset/limit to the region, grep/find/query to
  locate — don't read a file just to search it.
- Track steps with todo and record findings in scratchpad (durable within the
  run; context can be compressed).

inherit_system_prompt stays False: the full parent prompt carries orchestration
and spawn_agent sections that are irrelevant to a sub-agent and references
tools it lacks. The sub-agent keeps its own focused, self-contained prompt.

Co-Authored-By: Claude <noreply@anthropic.com>

navi_code: brief sub-agent on its restricted toolset + context_transfer ...

The sub-agent briefing section told the model to "give the sub-agent
everything it needs" but never said which tools the sub-agent lacks or how
context actually reaches it — so briefings could ask the sub-agent to use
memory, spawn further agents, or switch profiles, all of which it can't.

- context_transfer: write the context the sub-agent needs (files, snippets,
  how to verify) into the scratchpad `context_transfer` section before
  spawning — it is injected automatically; the sub-agent does not inherit
  short-term memory or conversation history.
- Restricted toolset: the sub-agent has todo/scratchpad/reflect/filesystem/
  code_exec/terminal/list_tools but NOT memory/switch_profile/spawn_agent/
  schedule_recall/manage_recall — brief it to use only what it has (e.g.
  record findings in scratchpad, not memory).

The profile list is not duplicated — spawn_agent's own description carries it.

Co-Authored-By: Claude <noreply@anthropic.com>

recall: carry self-instruction (message) on the recall_update wire ...

The recall card could show call_type/trigger_at but not the self-instruction
(additional_context_message) — it was absent from RecallUpdate.to_wire, so
the user couldn't see what future-self was about to do at the scheduled or
fired moment. Extend the wire payload.

- events.RecallUpdate: add `message` field; to_wire emits "message".
- scheduler._publish_recall_update: accept and forward `message`.
- Publish sites carry message=recall.additional_context_message:
  schedule_recall (scheduled) and orchestrator._finalize_recall
  (rescheduled / fired / cancelled). manage_recall cancel/skip omit it
  (no recall object handy; already visible in the prior scheduled card).
- TUI RecallRenderer: preview the message (first line + "(+N lines)",
  capped at 80) on a `msg:` body line when present.
- Tests: RecallUpdate.to_wire carries message (defaults None); renderer
  preview (single/multiline/truncate/empty) and scheduled-card renders it.

Co-Authored-By: Claude <noreply@anthropic.com>

navi_code: leverage context-org/planning tools in system prompt ...

Wire the KV-backed context/planning tools into the profile so the agent
uses them instead of relying on lossy conversation memory under compression.

- Workflow Plan: create a todo for non-trivial tasks.
- Workflow Test & verify: record the verification in the todo `validation`
  field when marking done (structural form of "never claim done without
  verification").
- Replace "Context drift recovery" with "Working state & memory":
  - todo: plan + verification tracking.
  - scratchpad: durable working memory across compression (sections
    goal/findings/errors/artifacts); read before final report.
  - Sub-agent handoff: write context to the scratchpad `context_transfer`
    section before spawn_agent — it is injected into the sub-agent
    automatically (the sub-agent does not inherit short-term memory).
  - schedule_recall: continue after the iteration limit, offload heavy
    work headlessly, poll builds/logs, chain multi-phase work.
  - reflect: selectively, before complex plans or when stuck (3 LLM calls).
  - memory: global cross-project facts, not a scratchpad/docs substitute.
  - Drift recovery folded into a closing paragraph aligned with the
    goal_anchoring machinery.

TUI visualization for schedule_recall is a follow-up task.

Co-Authored-By: Claude <noreply@anthropic.com>