* fix(kimi): resolve Kimi Code API 403 errors and User-Agent restrictions
Kimi Code subscription keys require a whitelisted coding-agent User-Agent to avoid access_terminated_error 403s. This adds User-Agent probing and caching for Kimi Code endpoints.
Co-authored-by: Cursor <cursoragent@cursor.com>
* fix(kimi): omit temperature for kimi-for-coding API calls
Kimi Code rejects any non-default temperature with HTTP 400, which broke deep research probes and low-temp LLM rounds.
Co-authored-by: Cursor <cursoragent@cursor.com>
---------
Co-authored-by: Cursor <cursoragent@cursor.com>
* fix(agent): don't let a materialized default budget defeat context scaling
#1230 scales agent_input_token_budget to the model's context window unless
the user explicitly set a budget, detected via is_setting_overridden(). But
the settings-save path materializes every DEFAULT_SETTINGS key into
settings.json (load_settings merges defaults; handlers persist the merged
dict), so the persisted default 6000 reads as "overridden" and the budget
code takes the min(6000, ctx) branch — silently re-capping long-context
models at 6000 for anyone who has ever saved a setting. This reintroduces
the exact regression #1170/#1230 set out to fix.
Add is_setting_customized() (saved value != default) and gate the scaling
on it instead of mere presence. A persisted default is not a user choice.
is_setting_overridden has exactly one consumer (this budget path), so the
change is contained. Tests cover the materialized-default regression, a
deliberately-chosen budget still being honoured, and the absent-key case.
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
* fix(agent): rework context-budget fix per review (#4122)
Address RaresKeY's review:
P2 (explicitness): is_setting_customized treated a saved value equal to the
default as "not explicit", which ALSO blocked a user from deliberately pinning
the default budget. Reframe the default value itself as the AUTO sentinel —
agent_input_token_budget == DEFAULT_BUDGET means "scale to the model's context
window", any other value is an explicit cap. A materialized default still reads
as auto (fixing the original regression), and any non-default value the user
chooses is now honoured. Drop the now-unused is_setting_customized helper.
P2 (fallback context): auto-scaling trusted get_context_length() even when it
returned only the bare DEFAULT_CONTEXT fallback (no endpoint-reported / known
window), over-allocating on self-hosted/proxy setups. Add get_context_length_known()
(also returns whether the window was actually discovered); the budget block
passes 0 when unknown so auto-scaling stays conservative instead of inflating to
an unproven window.
hard_max stays auto-only — a deliberate explicit budget wins (#1190); kept that
contract and answered the reviewer's question rather than silently reversing it.
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
* test(agent): lock the materialized-default budget regression (review on #4121)
Per WGlynn's review on the issue: add an end-to-end regression that saves an
UNRELATED setting (which makes the settings-save path materialize the budget
default into settings.json) and asserts the budget still auto-scales rather than
re-reading as an explicit 6000 cap — locking the exact reopening shut.
To make the test bite the production decision (not just re-derive it), extract
`budget_is_explicit()` into src/context_budget.py and use it from the agent loop.
It keys off value-vs-default (the default is the auto sentinel), NOT settings
presence — which is the whole point, since the save path materializes defaults.
Note: after this PR's rework, is_setting_overridden has ZERO production callers,
so the merged-dict materialization smell can't reach any setting through a
presence check today (WGlynn's durability concern).
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
* fix(agent): bind the budget context window to its own provenance (review #4122)
RaresKeY caught a correctness bug in the fallback-context guard: stream_agent_loop
kept only the `known` flag from get_context_length_known() and budgeted off the
passed-in `context_length`, which can come from a *different* lookup. Two failures:
- local endpoints are re-queried, so the passed value can be a stale DEFAULT_CONTEXT
fallback while the fresh probe proves the real (smaller) served context — we'd
scale off the stale value;
- callers that don't pass context_length (scheduled tasks, teacher escalation,
skill test runs, bg_monitor) were capped at 6000 even when a long window is
discoverable.
Extract budget_context_for_model() which returns the freshly-probed window when
known else 0, binding the flag to the value it proves; the agent loop uses it.
Regression tests cover the stale-fallback, no-arg-caller, and probe-error paths.
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
* docs(agent): fix stale budget comments + tighten to the contract (review #4122)
- settings.py: an explicit budget is clamped to the window only — hard_max is
auto-only (#1190); drop the incorrect "and to hard_max".
- is_setting_overridden docstring: drop the stale "adaptive budgets" example;
point value-sensitive callers at context_budget.budget_is_explicit.
- Tighten the budget-block comments to the contract (default = auto sentinel,
non-default = explicit cap, hard_max = auto-only ceiling).
Comment/docstring-only; no behaviour change.
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
* docs(agent): correct budget issue citations (#1190 → merged #1230/#1273)
The context-budget contract (auto-sentinel, explicit budgets honoured,
hard_max auto-only) merged via #1230 — #1190 was the earlier, closed,
superseded PR. Re-point the contract comments at #1230 (the live source,
already cited for the auto-sentinel two lines up in settings.py).
The configurable hard_max setting (`agent_input_token_hard_max`) was a
reviewer requirement first raised on #1190, omitted from the merged #1230,
and actually added in #1273 — credit #1273 for it and correct the test
comment's history (it previously implied this PR completed the requirement).
Comment/docstring-only; no behaviour change.
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
---------
Co-authored-by: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
Make src.youtube_handler a compatibility wrapper around services.youtube.youtube_handler so transcript state, URL parsing, and timeout behavior no longer diverge.
Lock the API key encryption key file to owner-only permissions on creation and when reading existing keys, with regression coverage for permissions and encryption roundtrip.
load_settings() already catches PermissionError, but load_features() caught only
FileNotFoundError/JSONDecodeError/ValueError. An existing-but-unreadable
data/features.json (e.g. root-owned after a deploy) therefore raised instead of
falling back to DEFAULT_FEATURES, taking down GET /api/auth/features and anything
that reads feature flags. Add PermissionError to the except tuple to match
load_settings().
Adds tests/test_load_features_permission_error.py.
Co-authored-by: Alexandre Teixeira <111787685+alteixeira20@users.noreply.github.com>
* feat(agent): workspace confinement via context-local binding + get_workspace tool
Bind the per-turn workspace once in execute_tool_block; the shared path
resolvers (_resolve_tool_path / _resolve_search_root) and the subprocess cwd
helper (agent_cwd) read it, so file tools + bash/python are confined centrally
and a new tool that uses the shared helpers cannot accidentally bypass it.
Adds the admin-gated /api/workspace/browse picker, a workspace pill + directory
modal (reusing existing modal/button CSS), the /workspace slash command, and a
get_workspace tool (replaces a system-prompt block). Confinement is OS-agnostic
(realpath/normcase/commonpath) and docker-safe (container paths, no host
assumptions). Reopens#2023.
* ux(workspace): clarify workspace is not a sandbox
Picker modal note + pill tooltip + get_workspace tool/output wording now state
plainly: read_file/write_file/edit_file/grep/glob/ls are confined to the folder,
but bash/python only start there (cwd) and are not sandboxed. Modal note reuses
the existing .muted class.
* fix(agent): treat an active workspace as file-work intent
A vague low-signal message (e.g. "look at the local project") matches no
domain keywords, so tool retrieval is skipped and only always-available tools
are offered — leaving the agent with no file access even though a workspace is
set. When a workspace is active, include the file/code tools (incl.
get_workspace) on low-signal turns so the agent can act on the folder.
Also requires the tool index (ChromaDB) to be reachable for normal retrieval;
that is an environment dependency, not part of this change.
* ux(workspace): hide pill + overflow entry in chat mode
Workspace only scopes the agent's file/shell tools, so the pill and the
overflow 'Workspace' entry are agent-only now — hidden in chat mode like the
bash toggle. Mode read from the DOM in syncWorkspaceIndicator; applyMode() is
called from the agent/chat setMode handler.
* prompt(tools): steer bash/python to defer to the dedicated file tools
bash/python schema descriptions (what native-tool-calling models read) were
bare and gave no steer, so models would do file ops via the shell (e.g. writing
SVG/HTML, which then dumps raw markup into the tool preview). Tell bash/python
in the schema + tool-index + prompt section to prefer read_file/write_file/
edit_file/grep/glob/ls and only be used for what those do not cover.
* prompt(tools): keep bash/python deferral generic (no hardcoded tool names)
Reference 'a dedicated tool' rather than listing read_file/write_file/grep/etc.
by name, so the guidance does not go stale if those tools are renamed.
* style(workspace): drop em-dashes from added code comments/strings
* ux(workspace): terser non-sandbox note in picker (no tool-name list)
* ux(workspace): mirror terse non-sandbox wording in pill tooltip
* chore: untrack local venv symlink (run-only, not part of the feature)
* prompt(workspace): keep get_workspace text generic (no hardcoded tool names)
* fix(agent): low-signal + workspace surfaces only read-only file tools
Intersect the files tool group with PLAN_MODE_READONLY_TOOLS so a vague message
in a workspace exposes read_file/grep/glob/ls/get_workspace for exploration, but
not write_file/edit_file/bash/python -- those wait for a request that actually
calls for them (RAG retrieval still adds them on a real ask).
* feat(workspace): cap browse listing at 500 dirs with a truncated hint
Mirror the filesystem_tools._CODENAV_MAX_HITS pattern with a module-local
_MAX_BROWSE_DIRS so a directory with thousands of children does not dump every
row into the picker; the response carries a truncated flag and the modal tells
the user to type a path to jump in.
* chore: untrack local venv symlink (run-only artifact)
* fix(workspace): vet the workspace root against the sensitive-path deny list at bind time
The in-workspace resolver deny-lists sensitive paths inside the workspace,
but the empty-path search root is the workspace itself, so a workspace of
~/.ssh could be listed via ls with no path. vet_workspace() (public, in
tool_execution next to the resolvers) rejects non-directories and sensitive
roots before the path is ever bound; chat_routes uses it instead of its
inline isdir check.
* fix(workspace): reject filesystem roots and stop showing rejected workspaces as active
Review findings from #3665:
P2: vet_workspace accepted / (and would accept drive/UNC roots), which makes
every absolute path 'inside' the workspace and collapses confinement into
host-wide file access. A root is its own dirname, so reject when
dirname(resolved) == resolved; the browse response now carries a selectable
flag and the picker disables 'Use this folder' on unselectable dirs.
P3: /workspace set stored any string client-side and the chat route silently
dropped rejected values, so the pill could claim a confinement that was not
in effect. New admin-gated /api/workspace/vet validates manual paths before
they persist (canonical path returned), and when a posted workspace is
rejected at send time the stream emits workspace_rejected so the client
clears the stored value and toasts instead of continuing silently.
* fix(workspace): check caller privilege before vetting the posted workspace
Review finding: /api/chat_stream called vet_workspace() on the posted value
for every caller and emitted workspace_rejected on failure, so a non-admin
who can chat but cannot use file/shell tools could distinguish existing
directories from missing/file/sensitive/root paths by whether the event
appeared. The resolution now lives in _resolve_request_workspace, which
drops the submitted value uniformly for non-admin callers, with no vetting
and no event, before the path ever touches the filesystem. Admin and
single-user behavior is unchanged. Test pins that valid and invalid paths
are indistinguishable for a non-admin and that vet_workspace is never
invoked for them.
Replace hardcoded [:2000] and [:4000] slicing with the shared _truncate
helper from tool_utils, which uses MAX_OUTPUT_CHARS and adds an explicit
truncation indicator when content is cut.
Scoped down from the original PR: only agent/tool-output display
behavior, no integrations.py changes.
Co-authored-by: michaelxer <michaelxer@users.noreply.github.com>
Co-authored-by: Alexandre Teixeira <111787685+alteixeira20@users.noreply.github.com>
* fix(llm): stop sending llama.cpp slot-affinity fields to cloud providers
_apply_local_cache_affinity adds session_id + cache_prompt for llama.cpp
KV-cache slot affinity (#2927), gated on _is_self_hosted_openai_compatible,
which treated any unknown OpenAI-compatible host as self-hosted. Strict
cloud providers added as custom endpoints (Mistral at api.mistral.ai)
reject unknown body fields, so every request failed with 422
extra_forbidden. Self-hosted now also requires the endpoint to resolve as
local via model_context.is_local_endpoint: loopback/private/tailscale
host, or endpoint kind explicitly configured as "local" (the escape hatch
for tunneled self-hosted servers). is_local_endpoint is promoted to a
public name since llm_core now shares it.
Fixes#3793
* test(llm): sweep cloud OpenAI-compatible hosts in affinity gating
Parametrized cases adapted from #3839 (credit: Shabablinchikow): deepseek,
x.ai, together, fireworks, and the Gemini OpenAI-compat endpoint must all
stay free of the llama.cpp extras, not just the Mistral host from #3793.
* fix(llm): narrow the Tailscale range to 100.64.0.0/10 in is_local_endpoint
Review finding on #3945: _PRIVATE_PREFIXES carried a bare "100." prefix,
treating all of 100.0.0.0/8 as local while Tailscale only uses the CGNAT
block 100.64.0.0/10. Public 100.x hosts (e.g. AWS ranges outside the
block) were classified local and still received the llama.cpp extras
this PR exists to keep away from strict providers. Match the narrowed
classification routes/model_routes.py already uses, with boundary tests
just below, inside, and just above the range.
_search_fts ran the FTS MATCH query, then looked up each hit's full row with its
own db.query(...).filter(id == message_id).first() inside a loop, so a search
returning N hits issued N extra SELECTs. Fetch all hit rows in a single IN(...)
query via _fetch_messages_by_id and reassemble results in hit (relevance) order.
Adds tests/test_session_search_batch_fetch.py asserting a single batched query
(and no query for empty input). Existing session-search tests stay green.
fire() and fire_and_forget() scheduled delivery with bare create_task()/
loop.create_task() and kept no reference. asyncio holds only a weak reference to
a task, so the GC could collect a delivery (or the fire() coroutine itself)
before it completed, silently dropping the webhook.
Track in-flight tasks in a set on the manager via a _spawn_tracked() helper that
holds a strong reference for the task's lifetime and discards it on completion
(add_done_callback), and route both schedule sites through it.
Adds tests/test_webhook_task_refs.py.
Anthropic removed the sampling parameters (temperature, top_p, top_k)
starting with Claude Opus 4.7 — sending temperature at all, even 0.0,
returns HTTP 400. _build_anthropic_payload sent it unconditionally, so
every native-Anthropic request to Opus 4.7/4.8 failed: the research probe
(ResearchHandler._probe_endpoint, temperature=0) aborted runs before they
started, and all DeepResearcher._llm calls 400'd.
Add _anthropic_rejects_temperature (version-gates opus-N-M >= (4,7)) and
omit temperature in the Anthropic builder for those models. Older Claude
models (Opus 4.6 and below, Sonnet/Haiku) keep temperature and the
existing [0,1] clamp.
The version gate is hardened against real-world model id shapes:
- a word-boundary anchor so a substring like `octopus-4-8` is not read
as Opus and stripped of temperature;
- a 1-2 digit minor cap so a dated id such as `claude-opus-4-20250514`
(Opus 4.0, listed in ANTHROPIC_MODELS) parses as major-only and keeps
temperature, while dated 4.7+ snapshots still match;
- a non-string guard so a non-string model can't raise AttributeError
(the previous builder never called .lower() on it).
Adds regression tests covering 4.7/4.8 omission, older/dated/legacy
retention, the substring overmatch, and non-string inputs.
Fixes#3065
Co-authored-by: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
_warmup_endpoints called model_discovery.get_endpoints(), which does not exist
on ModelDiscovery. It raised AttributeError on every startup and on every 60s
keepalive tick, was swallowed by the outer except, and pinged nothing, so the
cold-start prevention the loop exists for never ran.
Add ModelDiscovery.warmup_ping_urls(), which resolves the /models probe URLs
from the real discover_models() output, and call it from the warmup loop via
asyncio.to_thread (discovery does a blocking port scan, so keep it off the event
loop).
Adds tests/test_warmup_ping_urls.py: resolves /models URLs from discovered
items, honors the limit, degrades to [] on discovery failure, and documents that
get_endpoints never existed.
* fix: handle batch events format in manage_calendar tool
Models like deepseek-v4-flash emit batch events array instead of individual create_event calls. The tool defaulted to list_events (no action key), so events were never created despite the model confirming success.
- Add batch normalization in do_manage_calendar
- Map start/end objects to flat dtstart/dtend strings
- Add tests for both object and flat string formats
* fix: surface partial batch failures in manage_calendar
Partial failures were silently dropped - batches with mixed success/failure would report only created count with no error visibility.
- Return non-zero exit code for any failures
- Surface both created and failed counts in response
- Include first error message for debugging
- Add test for partial failure case
* chore: strip trailing whitespace in batch normalization block
* chore: strip whitespace-only blank lines in batch events test
The classify_events task pulled user memories to give the LLM personal context,
but read `m.content`, which the Memory ORM does not have (the column is `text`).
That raised AttributeError on the first row; the surrounding except swallowed it
and logged at debug, so the personal-context block was silently always empty and
events were classified without it.
Extract the rendering into `_memory_context_lines` (reads `text`, robust via
getattr, keeps the 200-char and 40-line caps) and raise the swallowed-exception
log to warning so a future schema mismatch is visible.
Adds tests/test_classify_events_memory_text.py for the field, truncation, blank
skipping, missing-attr robustness, and the line cap.
get_status() called get_avg_duration() unconditionally, and that helper globs
and JSON-parses every file under the research data dir. The SSE status stream
polls get_status() roughly once a second, so with a few saved reports each poll
re-read and re-parsed all of them, including for sessions that are not active
(the disk branch never even used the value).
Compute avg_duration only for active sessions and memoize it on the task entry,
so a long stream computes it once instead of on every poll. Behaviour is
unchanged: active streams still report avg_duration.
Adds tests/test_research_status_avg_duration.py: an inactive session does no
avg scan, and an active session computes it once across many polls.
* fix(security): don't grant tool access in the pre-setup window
owner_is_admin_or_single_user() returned True whenever auth was not
configured, which conflated two very different states:
- intentional single-user mode (operator set AUTH_ENABLED=false), and
- the pre-setup window (auth enabled, but no admin created yet).
In the second state, blocked_tools_for_owner() returned an empty set, so
server-execution tools (bash/python) and other admin-only tools were
ungated. The auth middleware already 401s /api/ requests pre-setup, but a
caller that bypasses it (trusted loopback / internal-tool path) could reach
those tools before setup completed.
Treat "not configured" as admin only when auth is intentionally disabled
(AUTH_ENABLED=false), mirroring the AUTH_ENABLED parsing in app.py and
core.middleware. Single-user mode is preserved; the pre-setup window is now
non-admin as defense-in-depth.
Adds regression tests for both states.
Fixes#3201
Supported by Claude Opus 4.8
* refactor(security): reuse _auth_disabled() instead of a duplicate helper
Addresses review on #3506: src/auth_helpers.py already has _auth_disabled()
with the identical AUTH_ENABLED parse. Drop the duplicate
_auth_intentionally_disabled() and call the existing helper via a lazy import
inside owner_is_admin_or_single_user (mirroring the lazy core.auth import) to
avoid any import cycle. Removes the now-unused `import os`. Behaviour and the
two regression tests are unchanged.
Supported by Claude Opus 4.8
---------
Co-authored-by: SurprisedDuck <288741682+SurprisedDuck@users.noreply.github.com>
* fix(chat): stabilize system prompt, sequence memory extraction, send stable session id to preserve KV cache
Fixes#2927. As diagnosed in the issue, three things in Odysseus's request
pattern actively destroyed local backends' (llama.cpp / LM Studio) KV-cache
continuity, forcing a full prompt re-evaluation (15-30s+) on every turn:
1. Dynamic content folded into the system prompt every turn. Both the chat
preface (ChatProcessor.build_context_preface) and the agent system prompt
(_build_system_prompt) injected current_datetime_prompt() — text that
changes every minute — directly into system-role messages, which llm_core
then concatenates into the single system message sent as the cached
prefix. Any byte difference there invalidates the entire cache. Moved this
to a new current_datetime_context_message() helper that returns a
standalone user-role message, inserted near the end of the array (right
before the latest user turn) instead of mixed into the system prompt. The
static system prefix (preset prompt + safety policy + agent base prompt)
now stays byte-identical across turns of the same session.
2. Memory/skill extraction side-requests competed with the main completion.
run_post_response_tasks fired extract_and_store / maybe_extract_skill via
asyncio.create_task — fire-and-forget coroutines that could overlap the
next turn's main request and steal llama.cpp's limited processing slots,
evicting the cached checkpoint. They're now queued through a new
_run_extraction_jobs_sequentially helper that waits for the session's
stream to go idle and runs the jobs strictly one at a time.
3. No stable session identifier was sent to local backends, so llama.cpp
assigned a new processing slot via LRU every turn ("session_id=<empty>
server-selected (LCP/LRU)"), losing slot affinity. Added
_apply_local_cache_affinity() in llm_core, which sets session_id and
cache_prompt: true on outgoing payloads — gated to self-hosted
OpenAI-compatible endpoints only (never api.openai.com or other cloud
providers, which reject unrecognized request fields with a 400). Threaded
session_id through stream_llm / llm_call_async / stream_agent_loop from
the existing Odysseus session id.
Tests in tests/test_kv_cache_invalidation_2927.py exercise the real payload-
assembly and scheduling code paths: byte-identical system prefix across two
turns of the same session (with a regression check that genuinely changed
instructions DO still change it), the dynamic time block landing as a
user-role message, extraction jobs waiting for the stream to go idle and
running sequentially, and the outgoing payload carrying a stable session_id
(same across turns of one session, different across sessions) only for
self-hosted endpoints. Updated tests/test_user_time.py for the new message
placement.
* fix(tests): accept owner= kwarg in normalize_model_id monkeypatch
The upstream normalize_model_id signature now takes an owner= keyword
argument, and chat_helpers.py passes owner=getattr(sess, "owner", None)
at the call site. Update the test stub lambda to **kwargs so it handles
the new argument without breaking, and update chat_helpers.py to forward
the owner parameter consistently.
---------
Co-authored-by: Alexandre Teixeira <111787685+alteixeira20@users.noreply.github.com>
* fix(integrations): truncate api_call JSON lists with sentinel instead of mid-string cut
* fix(integrations): avoid mutating response dict in-place on truncation
* fix(integrations): truncate dict responses and bound list sentinel overhead
- Dict path now walks keys in insertion order, adding them one at a time
while checking that the accumulated dict + _truncated marker fits within
the 12 000-char limit. Previously the marker was appended without removing
any content, so large dicts were not actually truncated.
- List path now subtracts the sentinel's serialised size (+ element-separator
padding) from the budget before binary-searching, so the final array
including the sentinel stays at or under the limit.
- Add regression tests: large-dict actually-truncated, small-dict pass-through,
and list-with-sentinel respects the size bound.
---------
Co-authored-by: Alexandre Teixeira <111787685+alteixeira20@users.noreply.github.com>
Providers like Moonshot (Kimi K2.5/K2.6) require the reasoning_content
field to be present on assistant tool-call messages in multi-turn
conversations. The sanitizer's allow-list was missing this field,
causing HTTP 400: 'thinking is enabled but reasoning_content is missing
in assistant tool call message at index N'.
Add reasoning_content to the allowed field set in
_sanitize_llm_messages and cover with regression tests.
Fixes#3118
Co-authored-by: michaelxer <michaelxer@users.noreply.github.com>
Co-authored-by: Alexandre Teixeira <111787685+alteixeira20@users.noreply.github.com>
#3322 renamed the loopback base to _INTERNAL_BASE, but a later Cookbook
commit reintroduced one call site using the old _COOKBOOK_BASE name,
raising NameError whenever the agent registers a model endpoint for a
running serve session.
Fixes#3669
ChatGPT's Codex API rejects any request that includes max_output_tokens,
returning HTTP 400 "Unsupported parameter: max_output_tokens". This caused
Deep Research to always fail during the endpoint probe when a ChatGPT
Subscription model was selected.
Remove the conditional that set payload["max_output_tokens"] in
_build_chatgpt_responses_payload(). The parameter is simply not sent.
Also update the two affected tests:
- Rename test_chatgpt_subscription_payload_uses_max_output_tokens →
test_chatgpt_subscription_payload_omits_max_output_tokens
- Rename test_chatgpt_subscription_payload_omits_empty_max_output_tokens →
test_chatgpt_subscription_payload_omits_max_output_tokens_when_zero
- Assert max_output_tokens is absent rather than present
Fixes#3650
* Add consolidated service health endpoint for degraded-state reporting
ROADMAP (High Priority) asks for "Better degraded-state reporting for
ChromaDB, SearXNG, email, ntfy, and provider probes." Until now there was no
single readout of which subsystems are actually working: /api/health is only a
liveness ping and each subsystem's signal lives in a different module, so a
misconfigured self-host install gives no consolidated picture.
This adds an admin-only GET /api/diagnostics/services endpoint backed by a new
src/service_health.py aggregator. Each subsystem reports a uniform
{name, status, detail, meta} where status is ok | degraded | down | disabled,
and the response rolls up an overall verdict (worst non-disabled status).
Probes are deliberately non-intrusive and safe to poll:
- ChromaDB: reads the .healthy flags on the RAG and memory vector stores.
- SearXNG: GET /healthz (2xx), falling back to the instance root (<500). No
search query is run.
- ntfy: GET the server's built-in /v1/health. No test notification is sent.
- email: short IMAP connect+logout per configured account (no credentials in
meta).
- providers: probe each enabled ModelEndpoint's model list (no api_key in meta).
Probe functions take their inputs as parameters and isolate the network call to
injectable callables, so they unit-test without touching the network (same
pattern as the merged provider-endpoint tests). Network probes run concurrently
off the event loop via asyncio.to_thread with bounded per-probe timeouts.
memory_vector is now passed into setup_diagnostics_routes (new optional param,
backward-compatible) so ChromaDB's vector-memory store can be reported too.
Tests: tests/test_service_health.py — 29 tests covering every status mapping
per subsystem, the overall rollup, and that no secrets leak into meta.
Verification:
python -m pytest tests/test_service_health.py -q # 29 passed
python -m py_compile src/service_health.py routes/diagnostics_routes.py app.py
python -m pytest tests/test_endpoint_resolver.py tests/test_provider_endpoints.py -q
Backend + tests only; an Admin/Settings UI badge that renders this endpoint is
a natural follow-up.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
* fix(diagnostics): bound service-health wall-clock and redact secrets
Addresses review on #964.
Blocker 1 — genuinely bounded wall-clock:
- providers_health and email_health now fan out per-item probes across a
bounded thread pool (_bounded_map) with a hard total budget (_FANOUT_BUDGET),
instead of probing endpoints/accounts sequentially. Stragglers are reported
as a controlled `timeout` and never block; the pool is shut down with
wait=False so the response returns on time regardless of endpoint/account
count.
- The IMAP connect path now honors the service-health budget: _imap_connect
gained a pass-through `timeout` param and the probe calls it with
_PROBE_TIMEOUT instead of the default 15s.
- collect_service_health runs the four network subsystems concurrently, each
under a per-subsystem deadline (_SUBSYSTEM_DEADLINE), with an overall
wait_for ceiling (_AGGREGATE_DEADLINE) as a backstop.
Blocker 2 — no secret/raw-error leakage in the response:
- _safe_url strips userinfo, query, and fragment from every URL surfaced in
meta (searxng instance, ntfy base, provider name fallback), keeping only
scheme/host/port/path.
- _classify_error maps every probe failure to a controlled category token
(timeout, connection_refused, dns_error, tls_error, network_error,
http_error, auth_or_protocol_error, …) — raw str(exception), which can embed
credentialed URLs or server text, is never returned.
Tests (tests/test_service_health.py, +tests/test_diagnostics_service_route.py):
- URL userinfo/query redaction for searxng/ntfy/providers.
- secret-bearing exception strings map to categories and don't leak.
- multiple slow providers/accounts stay bounded (single + 25-endpoint cases).
- subsystems run concurrently; aggregate deadline yields a controlled result.
- route-level unauthenticated (401) / non-admin (403) / admin (200) coverage.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
* test(diagnostics): isolate route tests so they don't leak module globals
The new route tests replaced src.service_health.collect_service_health and
routes.diagnostics_routes.require_admin via direct assignment, which persisted
for the rest of the pytest session. In CI's full alphabetical run that fake
collector (returning services=[]) leaked into the later collect_service_health
tests and failed them. Switch to monkeypatch.setattr so both are restored after
each test. No production code change.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
---------
Co-authored-by: Claude Opus 4.8 <noreply@anthropic.com>
Co-authored-by: Alexandre Teixeira <111787685+alteixeira20@users.noreply.github.com>
* refactor(tools): implement strict cohesive class coordinator pattern per #2917
* test: update edit_file tests to use EditFileTool class
* fix(tools): restore tool_policy param and security backstop in coordinator
* refactor(tools): migrate domain tools to agent_tools package per #2917
* test: update test imports for new agent_tools package
* fix: resolve circular import between tool_execution and agent_tools
* fix: remove leftover git conflict markers
* fix(tools): resolve pytest failure and document _apply method
* fix(tools): clean up whitespace and remove dead _tool_python helper
---------
Co-authored-by: Alexandre Teixeira <111787685+alteixeira20@users.noreply.github.com>
* docs: add implementation plan for fixing chat context drifting (#135)
* fix: make Session.history immutable + fix {}.history crash
- Session.history now exposes a COPY of the internal _history list
- add_message() replaces history with a fresh copy each time
- get_context_messages() derives from _history directly
- replace_messages() updates both _history and history
- truncate_messages() updates both _history and history
- _persist_message() line 207: fixed {}.history fallback crash
- Added 11 tests for session isolation and edge cases
Addresses #135 root cause #1: shared mutable references
* fix: task scheduler uses SessionManager methods instead of overwriting sessions
- Added ensure_task_session() to SessionManager (checks cache first)
- Task scheduler now uses ensure_task_session() instead of direct dict assignment
- Task scheduler now uses SessionManager.add_message() for message persistence
- Removed direct sess_obj.history.append() that was silently losing data
Addresses #135 root causes #2 and #3
* fix: add age guard to cleanup_empty_sessions — don't delete sessions <1h old
Prevents the cleanup task from deleting sessions that were just created
and haven't received any messages yet (message_count == 0).
Addresses #135 root cause #5
* test: comprehensive session isolation tests (10/10 passing)
* refactor: consolidate _session_manager into singleton pattern
- Added set_session_manager_instance / get_session_manager_instance to core/models
- kept backward-compat aliases (set_session_manager, get_session_manager)
- session_manager.py re-exports the singleton functions
- ai_interaction.set_session_manager now syncs with the core singleton
- context_compactor uses get_session_manager_instance() instead of getattr hack
- app.py initializes the singleton once
Addresses #135 root cause #4: fragile global wiring
* test: add concurrent session isolation integration tests
Verifies:
- Concurrent add_message to different sessions doesn't cross-contaminate
- Rapid parallel writes maintain isolation
- Read-write concurrent access is safe
All 3 async tests pass, proving the immutable history fix works under concurrency
* fix: pre-import core.models in conftest to prevent test pollution
test_agent_loop.py stubs sys.modules['core.models'] = MagicMock() at
module level during collection. Any test collected after it imports
Session as a MagicMock. Pre-importing core.models in conftest.py
before test_agent_loop.py's module-level code runs prevents this.
* fix: make .history authoritative mutable list, address PR review
Per review feedback: keep .history as the authoritative mutable list so
existing code doing .history.pop(), .history = [...], etc. still works.
Fix the cross-contamination bug by ensuring __post_init__() gives each
Session its OWN unique history list (never shared).
Changes:
- core/models.py: .history IS the authoritative list. _history aliases it.
Each Session gets its own list in __post_init__.
- core/session_manager.py: add_message() delegates to Session.add_message()
instead of appending directly — no double-append, single source of truth.
- tests/test_session_manager.py: updated test to reflect that .history
references see new messages (same list, not a snapshot).
- docs/plans/2026-06-01-fix-chat-context-drifting.md: removed (not for
shipping — useful design context but too much process/doc to ship).
All 272 tests pass (3 pre-existing failures unrelated).
* Fix session manager message persistence
* Fix session history alias regressions
* Fix session history aliasing and task delivery
Achilleas90