- Remove dead _lmstudio_loaded_context attribute from run_agent.py (set
but never read — the loaded context is pushed to context_compressor.update_model
which is the actual consumer)
- Cache empty reasoning options with 60s TTL to avoid per-turn HTTP probe
for non-reasoning LM Studio models. Non-empty results cached permanently.
- Extract _lmstudio_server_root(), _lmstudio_request_headers(), and
_lmstudio_fetch_raw_models() shared helpers in models.py — eliminates
URL-strip + auth-header + HTTP-call duplication across probe_lmstudio_models,
ensure_lmstudio_model_loaded, and lmstudio_model_reasoning_options
- Revert runtime_provider.py base_url precedence change: preserve the
established contract (saved config.base_url > env var > default) for all
api_key providers
- Remove unnecessary config version bump 22→23
- Fix TUI test: relax target_model assertion to avoid module-cache flake
- AUTHOR_MAP: added rugved@lmstudio.ai → rugvedS07
Anyone who ran hermes between Apr 15 (42aeb4ec) and Apr 22 (a7d78d3b)
has schema_version=7 from the pre-renumber api_call_count migration.
When a7d78d3b inserted reasoning_content as the new v7 and pushed
api_call_count to v8, the 'if current_version < 7' gate was already
false for those users, so reasoning_content was never created —
sqlite3.OperationalError: no such column: reasoning_content on any
/continue or /resume touching assistant replays.
Replaces the version-gated ADD COLUMN chain with _reconcile_columns():
on every startup, parse SCHEMA_SQL via an in-memory SQLite and diff
against PRAGMA table_info; ALTER TABLE ADD COLUMN for anything missing.
Follows the Beets / sqlite-utils pattern — SCHEMA_SQL becomes the single
source of truth for declared columns. Self-healing and idempotent.
v10 trigram FTS backfill is retained in a version-gated block — that
migration isn't a column add, it inserts existing message rows into
the new FTS virtual table, so reconciliation can't express it.
schema_version is also kept for future row-data migrations.
Salvaged from #14097 (@kshitijk4poor) onto current main; v10 trigram
preservation and the v9 codex_message_items column (stale-missed by
the original branch) are covered automatically by reconciliation.
Tests:
- Regression: DB at old v7 with api_call_count but no reasoning_content
gets the column on open
- Idempotency: reopening the same DB is a no-op
- Structural invariant: every SCHEMA_SQL column is in the live DB
- Existing v2 migration test still passes
- E2E verified against fresh / v1 / old-v7 / v9 DBs, plus v10 trigram
backfill preserved
- config.py: remove dead ENV_VARS_BY_VERSION[17] entry (current _config_version
is 22, so all users are past version 17 and would never be prompted for
GMI_API_KEY on upgrade — consistent with how arcee was added)
- auxiliary_client.py: use google/gemini-3.1-flash-lite-preview as GMI aux
model instead of anthropic/claude-opus-4.6 (matches cheap fast-model pattern
used by all other providers: zai→glm-4.5-flash, kimi→kimi-k2-turbo-preview,
stepfun→step-3.5-flash, kilocode→google/gemini-3-flash-preview)
- test_gmi_provider.py: fix malformed write_text() call in doctor test
(was: write_text("GMI_API_KEY=*** encoding="utf-8") → missing closing quote,
wrote literal string 'GMI_API_KEY=*** encoding=' to .env file)
- test_gmi_provider.py + test_auxiliary_client.py: update aux model assertions
to match new cheaper default
- docs/integrations/providers.md: add 'gmi' to inline 'Supported providers'
fallback list (was only in the table, not the inline list at line ~1181)
- docs/reference/cli-commands.md: add 'gmi' to --provider choices list
When switching models on a custom endpoint (ollama-launch):
- Same-provider switches no longer re-resolve credentials (fixes base_url
being lost for 'custom' provider on subsequent switches)
- Named providers (ollama-launch) are resolved via user_providers so
switch_model can find their base_url from config
- Models not in the /v1/models probe but present in the user's saved
provider config are accepted with a warning instead of rejected
- CLI /model and TUI /model both pass user_providers/custom_providers
to switch_model so the config model list is available for validation
Closes#15088
The Codex Responses API rejects input_text inside assistant messages —
only output_text and refusal are valid content types for assistant role.
_chat_content_to_responses_parts() previously hardcoded all text content
to input_text regardless of the message role. When an assistant message
had list-format content (multimodal or structured), this produced invalid
input_text parts that the API rejected with:
Invalid value: 'input_text'. Supported values are: 'output_text' and 'refusal'.
Fix: add a role parameter to _chat_content_to_responses_parts() that
selects output_text for assistant messages and input_text for user
messages. Thread this through _chat_messages_to_responses_input() and
_preflight_codex_input_items().
Fixes#15687
When a user sends /stop during a streaming API call, the outer poll loop
detects _interrupt_requested and closes the HTTP connection. However, the
inner _call() thread catches the connection error and enters its retry
loop — opening a FRESH connection without checking the interrupt flag.
On slow providers like ollama-cloud, each retry attempt blocks for the
full stream-read timeout (120s+). With 3 retry attempts this caused
510+ second delays between /stop and actual response — the agent appeared
completely unresponsive despite the stop being acknowledged.
Fix: add an _interrupt_requested check at the top of the streaming retry
loop so the agent exits immediately instead of retrying.
Also fix log truncation: all session key logging in gateway/run.py used
[:20] or [:30] slices, which truncated 'agent:main:telegram:dm:5690190437'
(33 chars) to 'agent:main:telegram:' — losing the identifying chat type
and user ID. Replace with full keys to make logs debuggable.
Reported by user Sidharth Pulipaka via Telegram on ollama-cloud provider.
_check_compression_model_feasibility calls get_model_context_length
without provider=, so Codex OAuth users get 1,050,000 (from models.dev
for 'openai') instead of the actual 272,000 limit. This happens because
_infer_provider_from_url maps chatgpt.com → 'openai' (not 'openai-codex'),
skipping the Codex-specific resolution branch entirely.
Result: compression threshold set at 85% of 1.05M = 892K — conversations
never trigger compression, the context grows unbounded, and when gateway
hygiene eventually forces compression, the Codex endpoint drops the
oversized streaming request ('peer closed connection without sending
complete message body').
Fix: forward self.provider to get_model_context_length so provider-
specific resolution branches (Codex OAuth 272K, Copilot live /models,
Nous suffix-match) fire correctly.
Reported by user on GPT 5.5 via Codex OAuth Pro (paste.rs/vsra3).
Xiaomi's API (api.xiaomimimo.com) requires lowercase model IDs like
"mimo-v2.5-pro" but rejects mixed-case names like "MiMo-V2.5-Pro"
that users copy from marketing docs or the ProviderEntry description.
Add _LOWERCASE_MODEL_PROVIDERS set and apply .lower() to model names
for providers in this set (currently just xiaomi) after stripping the
provider prefix. This ensures any case variant in config.yaml is
normalized before hitting the API.
Other providers (minimax, zai, etc.) are NOT affected — their APIs
accept mixed case (e.g. MiniMax-M2.7).
Commit 43de1ca8 removed the _nr_to_assistant_message shim in favor of
duck-typed properties on the ToolCall dataclass. However, the
extra_content property (which carries the Gemini thought_signature) was
omitted from the ToolCall definition. This caused _build_assistant_message
to silently drop the signature via getattr(tc, 'extra_content', None)
returning None, leading to HTTP 400 errors on subsequent turns for all
Gemini 3 thinking models.
Add the extra_content property to ToolCall (matching the existing
call_id and response_item_id pattern) so the thought_signature round-trips
correctly through the transport → agent loop → API replay path.
Credit to @celttechie for identifying the root cause and providing the fix.
Closes#14488
Three bugs fixed in model alias resolution:
1. resolve_alias() returned the FIRST catalog match with no version
preference. '/model mimo' picked mimo-v2-omni (index 0 in dict)
instead of mimo-v2.5-pro. Now collects all prefix matches, sorts
by version descending with pro/max ranked above bare names, and
returns the highest.
2. models.dev registry missing newly added models (e.g. v2.5 for
native xiaomi). resolve_alias() now merges static _PROVIDER_MODELS
entries into the catalog so models resolve immediately without
waiting for models.dev to sync.
3. hermes model picker showed only models.dev results (3 xiaomi models),
hiding curated entries (5 total). The picker now merges curated
models into the models.dev list so all models appear.
Also fixes a trailing-dot float parsing edge case in _model_sort_key
where '5.4.' failed float() and multi-dot versions like '5.4.1'
weren't parsed correctly.
## Merged
Adds MiMo v2.5-pro and v2.5 support to Xiaomi native provider, OpenCode Go, and setup wizard.
### Changes
- Context lengths: added v2.5-pro (1M) and v2.5 (1M), corrected existing MiMo entries to exact values (262144)
- Provider lists: xiaomi, opencode-go, setup wizard
- Vision: upgraded from mimo-v2-omni to mimo-v2.5 (omnimodal)
- Config description updated for XIAOMI_API_KEY
- Tests updated for new vision model preference
### Verification
- 4322 tests passed, 0 new regressions
- Live API tested on Xiaomi portal: basic, reasoning, tool calling, multi-tool, file ops, system prompt, vision — all pass
- Self-review found and fixed 2 issues (redundant vision check, stale HuggingFace context length)
NormalizedResponse and ToolCall now have backward-compat properties
so the agent loop can read them directly without the shim:
ToolCall: .type, .function (returns self), .call_id, .response_item_id
NormalizedResponse: .reasoning_content, .reasoning_details,
.codex_reasoning_items
This eliminates the 35-line shim and its 4 call sites in run_agent.py.
Also changes flush_memories guard from hasattr(response, 'choices')
to self.api_mode in ('chat_completions', 'bedrock_converse') so it
works with raw boto3 dicts too.
WS1 items 3+4 of Cycle 2 (#14418).
3-layer chain (transport → v2 → v1) was collapsed to 2-layer in PR 7.
This collapses the remaining 2-layer (transport → v1 → NR mapping in
transport) to 1-layer: v1 now returns NormalizedResponse directly.
Before: adapter returns (SimpleNamespace, finish_reason) tuple,
transport unpacks and maps to NormalizedResponse (22 lines).
After: adapter returns NormalizedResponse, transport is a
1-line passthrough.
Also updates ToolCall construction — adapter now creates ToolCall
dataclass directly instead of SimpleNamespace(id, type, function).
WS1 item 1 of Cycle 2 (#14418).
Replace direct normalize_anthropic_response() call in
_AnthropicCompletionsAdapter.create() with
AnthropicTransport.normalize_response() via get_transport().
Before: auxiliary_client called adapter v1 directly, bypassing
the transport layer entirely.
After: auxiliary_client → get_transport('anthropic_messages') →
transport.normalize_response() → adapter v1 → NormalizedResponse.
The adapter v1 function (normalize_anthropic_response) now has
zero callers outside agent/anthropic_adapter.py and the transport.
This unblocks collapsing v1 to return NormalizedResponse directly
in a follow-up (the remaining 2-layer chain becomes 1-layer).
WS1 item 2 of Cycle 2 (#14418).
Consolidate 4 per-transport lazy singleton helpers (_get_anthropic_transport,
_get_codex_transport, _get_chat_completions_transport, _get_bedrock_transport)
into one generic _get_transport(api_mode) with a shared dict cache.
Collapse the 65-line main normalize block (3 api_mode branches, each with
its own SimpleNamespace shim) into 7 lines: one _get_transport() call +
one _nr_to_assistant_message() shared shim. The shim extracts provider_data
fields (codex_reasoning_items, reasoning_details, call_id, response_item_id)
into the SimpleNamespace shape downstream code expects.
Wire chat_completions and bedrock_converse normalize through their transports
for the first time — these were previously falling into the raw
response.choices[0].message else branch.
Remove 8 dead codex adapter imports that have zero callers after PRs 1-6.
Transport lifecycle improvements:
- Eagerly warm transport cache at __init__ (surfaces import errors early)
- Invalidate transport cache on api_mode change (switch_model, fallback
activation, fallback restore, transport recovery) — prevents stale
transport after mid-session provider switch
run_agent.py: -32 net lines (11,988 -> 11,956).
PR 7 of the provider transport refactor.
Upgrades agent-browser from 0.13.0 to 0.26.0, picking up 13 releases of
daemon reliability fixes:
- Daemon hang on Linux from waitpid(-1) race in SIGCHLD handler (#1098)
- Chrome killed after ~10s idle due to PR_SET_PDEATHSIG thread tracking (#1157)
- Orphaned Chrome processes via process-group kill on shutdown (#1137)
- Stale daemon after upgrade via .version sidecar and auto-restart (#1134)
- Idle timeout not firing (sleep future recreated each loop) (#1110)
- Navigation hanging on lifecycle events that never fire (#1059, #1092)
- CDP attach hang on Chrome 144+ (#1133)
- Windows daemon TCP bind with Hyper-V port conflicts (#1041)
- Shadow DOM traversal in accessibility tree snapshots
- doctor command for user self-diagnosis
Also wires AGENT_BROWSER_IDLE_TIMEOUT_MS into the browser subprocess
environment so the daemon self-terminates after our configured inactivity
timeout (default 300s). This is the daemon-side counterpart to the
Python-side inactivity reaper — the daemon kills itself and its Chrome
children when no commands arrive, preventing orphan accumulation even
when the Python process dies without running atexit handlers.
Addresses #7343 (daemon socket hangs, shadow DOM) and #13793 (orphan
accumulation from force-killed sessions).
Adds security.allow_private_urls / HERMES_ALLOW_PRIVATE_URLS toggle so
users on OpenWrt routers, TUN-mode proxies (Clash/Mihomo/Sing-box),
corporate split-tunnel VPNs, and Tailscale networks — where DNS resolves
public domains to 198.18.0.0/15 or 100.64.0.0/10 — can use web_extract,
browser, vision URL fetching, and gateway media downloads.
Single toggle in tools/url_safety.py; all 23 is_safe_url() call sites
inherit automatically. Cached for process lifetime.
Cloud metadata endpoints stay ALWAYS blocked regardless of the toggle:
169.254.169.254 (AWS/GCP/Azure/DO/Oracle), 169.254.170.2 (AWS ECS task
IAM creds), 169.254.169.253 (Azure IMDS wire server), 100.100.100.200
(Alibaba), fd00:ec2::254 (AWS IPv6), the entire 169.254.0.0/16
link-local range, and the metadata.google.internal / metadata.goog
hostnames (checked pre-DNS so they can't be bypassed on networks where
those names resolve to local IPs).
Supersedes #3779 (narrower HERMES_ALLOW_RFC2544 for the same class of
users).
Co-authored-by: kshitijk4poor <82637225+kshitijk4poor@users.noreply.github.com>
Drop rebased test assumptions about theme-mode helpers removed on main and keep the status bar skin integration aligned with the current skin engine model.
Route prompt_toolkit status bar colors through the skin engine so /skin updates the status bar alongside the rest of the interactive TUI.
Add regression coverage for the new status bar style override keys and CLI style composition.
These thin wrappers around _capture_log_snapshot had zero production
callers after the snapshot refactor — run_debug_share uses snapshots
directly and collect_debug_report captures internally. The wrappers
also caused a performance regression: _read_log_tail read up to 512KB
and built full_text just to return tail_text.
Remove both wrappers and migrate TestReadFullLog → TestCaptureLogSnapshot
to test _capture_log_snapshot directly. Same coverage, tests the real
API instead of dead indirection.
Add missing AUTHOR_MAP entry for taosiyuan163 whose truncation boundary
fix was adapted into _capture_log_snapshot().
Add regression tests proving: line-boundary truncation keeps the full
first line, mid-line truncation correctly drops the partial fragment.
Adapt the byte-boundary-safe truncation fix from PR #14040 by
taosiyuan163 into the new _capture_log_snapshot() code path: when
the truncation cut lands exactly on a line boundary, keep the first
retained line instead of unconditionally dropping it.
Also add a 2x max_bytes safety cap to the backward-reading loop to
prevent unbounded memory consumption when log files contain very long
lines (e.g. JSON blobs) with few newlines.
Based on #14040 by @taosiyuan163.
Adds _reactions_enabled() gating to match Discord (DISCORD_REACTIONS) and
Telegram (TELEGRAM_REACTIONS) pattern. Defaults to true to preserve existing
behavior. Gates at three levels:
- _handle_slack_message: skips _reacting_message_ids registration
- on_processing_start: early return
- on_processing_complete: early return
Also adds config.yaml bridge (slack.reactions) and two new tests.
Follow-up to the cherry-picked PR #13897 fix. Three issues found:
1. CRITICAL: The thinking block synthesised from reasoning_content was
immediately stripped by the third-party signature management code
(Kimi is classified as _is_third_party_anthropic_endpoint). Added a
Kimi-specific carve-out that preserves unsigned thinking blocks while
still stripping Anthropic-signed blocks Kimi can't validate.
2. Empty-string reasoning_content was silently dropped because the
truthiness check ('if reasoning_content and ...') evaluates to False
for ''. Changed to 'isinstance(reasoning_content, str)' so the
tier-3 fallback from _copy_reasoning_content_for_api (which injects
'' for Kimi tool-call messages with no reasoning) actually produces
a thinking block.
3. The thinking block was appended AFTER tool_use blocks. Anthropic
protocol requires thinking -> text -> tool_use ordering. Changed to
blocks.insert(0, ...) to prepend.
Adds schema v7 'api_call_count' column. run_agent.py increments it by 1
per LLM API call, web_server analytics SQL aggregates it, frontend uses
the real counter instead of summing sessions.
The 'API Calls' card on the analytics dashboard previously displayed
COUNT(*) from the sessions table — the number of conversations, not
LLM requests. Each session makes 10-90 API calls through the tool loop,
so the reported number was ~30x lower than real.
Salvaged from PR #10140 (@kshitijk4poor). The cache-token accuracy
portions of the original PR were deferred — per-provider analytics is
the better path there, since cache_write_tokens and actual_cost_usd
are only reliably available from a subset of providers (Anthropic
native, Codex Responses, OpenRouter with usage.include).
Tests:
- schema_version v7 assertion
- migration v2 -> v7 adds api_call_count column with default 0
- update_token_counts increments api_call_count by provided delta
- absolute=True sets api_call_count directly
- /api/analytics/usage exposes total_api_calls in totals
The transport refactor (PRs #13862 ff.) added agent/transports/ as a
sub-package but the setuptools packages.find include list only had
"agent" (top-level files), not "agent.*" (sub-packages).
pip install / Nix builds therefore ship run_agent.py (which now imports
from agent.transports on every API call) but omit the transports
directory entirely, causing:
ModuleNotFoundError: No module named 'agent.transports'
on every LLM call for packaged installs.
Adds "agent.*" to match the existing pattern used by tools, gateway,
tui_gateway, and plugins.
Fourth and final transport — completes the transport layer with all four
api_modes covered. Wraps agent/bedrock_adapter.py behind the ProviderTransport
ABC, handles both raw boto3 dicts and already-normalized SimpleNamespace.
Wires all transport methods to production paths in run_agent.py:
- build_kwargs: _build_api_kwargs bedrock branch
- validate_response: response validation, new bedrock_converse branch
- finish_reason: new bedrock_converse branch in finish_reason extraction
Based on PR #13467 by @kshitijk4poor, with one adjustment: the main normalize
loop does NOT add a bedrock_converse branch to invoke normalize_response on
the already-normalized response. Bedrock's normalize_converse_response runs
at the dispatch site (run_agent.py:5189), so the response already has the
OpenAI-compatible .choices[0].message shape by the time the main loop sees
it. Falling through to the chat_completions else branch is correct and
sidesteps a redundant NormalizedResponse rebuild.
Transport coverage — complete:
| api_mode | Transport | build_kwargs | normalize | validate |
|--------------------|--------------------------|:------------:|:---------:|:--------:|
| anthropic_messages | AnthropicTransport | ✅ | ✅ | ✅ |
| codex_responses | ResponsesApiTransport | ✅ | ✅ | ✅ |
| chat_completions | ChatCompletionsTransport | ✅ | ✅ | ✅ |
| bedrock_converse | BedrockTransport | ✅ | ✅ | ✅ |
17 new BedrockTransport tests pass. 117 transport tests total pass.
160 bedrock/converse tests across tests/agent/ pass. Full tests/run_agent/
targeted suite passes (885/885 + 15 skipped; the 1 remaining failure is the
pre-existing test_concurrent_interrupt flake on origin/main).
Third concrete transport — handles the default 'chat_completions' api_mode used
by ~16 OpenAI-compatible providers (OpenRouter, Nous, NVIDIA, Qwen, Ollama,
DeepSeek, xAI, Kimi, custom, etc.). Wires build_kwargs + validate_response to
production paths.
Based on PR #13447 by @kshitijk4poor, with fixes:
- Preserve tool_call.extra_content (Gemini thought_signature) via
ToolCall.provider_data — the original shim stripped it, causing 400 errors
on multi-turn Gemini 3 thinking requests.
- Preserve reasoning_content distinctly from reasoning (DeepSeek/Moonshot) so
the thinking-prefill retry check (_has_structured) still triggers.
- Port Kimi/Moonshot quirks (32000 max_tokens, top-level reasoning_effort,
extra_body.thinking) that landed on main after the original PR was opened.
- Keep _qwen_prepare_chat_messages_inplace alive and call it through the
transport when sanitization already deepcopied (avoids a second deepcopy).
- Skip the back-compat SimpleNamespace shim in the main normalize loop — for
chat_completions, response.choices[0].message is already the right shape
with .content/.tool_calls/.reasoning/.reasoning_content/.reasoning_details
and per-tool-call .extra_content from the OpenAI SDK.
run_agent.py: -239 lines in _build_api_kwargs default branch extracted to the
transport. build_kwargs now owns: codex-field sanitization, Qwen portal prep,
developer role swap, provider preferences, max_tokens resolution (ephemeral >
user > NVIDIA 16384 > Qwen 65536 > Kimi 32000 > anthropic_max_output), Kimi
reasoning_effort + extra_body.thinking, OpenRouter/Nous/GitHub reasoning,
Nous product attribution tags, Ollama num_ctx, custom-provider think=false,
Qwen vl_high_resolution_images, request_overrides.
39 new transport tests (8 build_kwargs, 5 Kimi, 4 validate, 4 normalize
including extra_content regression, 3 cache stats, 3 basic). Tests/run_agent/
targeted suite passes (885/885 + 15 skipped; the 1 remaining failure is the
test_concurrent_interrupt flake present on origin/main).
Add ResponsesApiTransport wrapping codex_responses_adapter.py behind the
ProviderTransport ABC. Auto-registered via _discover_transports().
Wire ALL Codex transport methods to production paths in run_agent.py:
- build_kwargs: main _build_api_kwargs codex branch (50 lines extracted)
- normalize_response: main loop + flush + summary + retry (4 sites)
- convert_tools: memory flush tool override
- convert_messages: called internally via build_kwargs
- validate_response: response validation gate
- preflight_kwargs: request sanitization (2 sites)
Remove 7 dead legacy wrappers from AIAgent (_responses_tools,
_chat_messages_to_responses_input, _normalize_codex_response,
_preflight_codex_api_kwargs, _preflight_codex_input_items,
_extract_responses_message_text, _extract_responses_reasoning_text).
Keep 3 ID manipulation methods still used by _build_assistant_message.
Update 18 test call sites across 3 test files to call adapter functions
directly instead of through deleted AIAgent wrappers.
24 new tests. 343 codex/responses/transport tests pass (0 failures).
PR 4 of the provider transport refactor.
Medium fixes:
- textInput.tsx: prevent silent data loss when async paste resolves
after user types — fall back to raw text insert at current cursor
instead of dropping the content entirely
- useComposerState.ts: tighten looksLikeDroppedPath to require a
second '/' or '.' for bare absolute paths, avoiding unnecessary
RPC round-trips for pasted text like /api or /help
- useComposerState.ts: add cross-reference comment linking to the
canonical _detect_file_drop() in cli.py
- osc52.ts: add 500ms timeout via Promise.race so terminals that
do not support OSC52 clipboard queries cannot hang paste
Low fixes:
- terminalSetup.ts: export isRemoteShellSession and reuse in
terminalParity.ts and useComposerState.ts (was inlined 3 times)
- useComposerState.ts: extract insertAtCursor helper, replacing 3
copies of the lead/tail spacing logic
- useComposerState.ts: remove redundant gw from handleTextPaste
useCallback dependency array
- terminalSetup.test.ts: add EACCES (read-only keybindings.json)
and unterminated block comment test coverage
Fixes from OutThisLife review:
1. Restore Linux Alt+Enter newline: textInput.tsx now uses
k.shift || (isMac ? isActionMod(k) : k.meta) so Alt+Enter
inserts a newline on Linux (was broken by isMac guard).
2. Fix image.attach response type: useComposerState.ts now uses
ImageAttachResponse (which already has remainder) instead of
InputDetectDropResponse with intersection.
3. Expand looksLikeDroppedPath test coverage with edge cases for
image extensions, file:// URIs, spaces, empty input, and
non-file URLs.
4. Make terminalParity.test.ts hermetic: terminalParityHints() now
accepts optional fileOps/homeDir and passes them through to
shouldPromptForTerminalSetup(), so tests inject mock readFile
instead of hitting the real filesystem.
Fixes from Copilot inline review:
5. Remove unused options.now parameter from configureTerminalKeybindings.
6. Replace naive stripJsonComments (full-line // only) with a proper
JSONC stripper that handles inline // comments, block comments,
trailing commas, and preserves comment-like sequences in strings.
7. Move backupFile() call from immediately after read to right before
write - backups are only created when changes will actually be
written, not on every /terminal-setup invocation.
Ports agent/account_usage.py and its tests from the original PR #2486
branch. Defines AccountUsageSnapshot / AccountUsageWindow dataclasses,
a shared renderer, and provider-specific fetchers for OpenAI Codex
(wham/usage), Anthropic OAuth (oauth/usage), and OpenRouter (/credits
and /key). Wiring into /usage lands in a follow-up salvage commit.
Authored-by: kshitijk4poor <82637225+kshitijk4poor@users.noreply.github.com>
Add agent/transports/types.py with three shared dataclasses:
- NormalizedResponse: content, tool_calls, finish_reason, reasoning, usage, provider_data
- ToolCall: id, name, arguments, provider_data (per-tool-call protocol metadata)
- Usage: prompt_tokens, completion_tokens, total_tokens, cached_tokens
Add normalize_anthropic_response_v2() to anthropic_adapter.py — wraps the
existing v1 function and maps its output to NormalizedResponse. One call site
in run_agent.py (the main normalize branch) uses v2 with a back-compat shim
to SimpleNamespace for downstream code.
No ABC, no registry, no streaming, no client lifecycle. Those land in PR 3
with the first concrete transport (AnthropicTransport).
46 new tests:
- test_types.py: dataclass construction, build_tool_call, map_finish_reason
- test_anthropic_normalize_v2.py: v1-vs-v2 regression tests (text, tools,
thinking, mixed, stop reasons, mcp prefix stripping, edge cases)
Part of the provider transport refactor (PR 2 of 9).
Extract 12 Codex Responses API format-conversion and normalization functions
from run_agent.py into agent/codex_responses_adapter.py, following the
existing pattern of anthropic_adapter.py and bedrock_adapter.py.
run_agent.py: 12,550 → 11,865 lines (-685 lines)
Functions moved:
- _chat_content_to_responses_parts (multimodal content conversion)
- _summarize_user_message_for_log (multimodal message logging)
- _deterministic_call_id (cache-safe fallback IDs)
- _split_responses_tool_id (composite ID splitting)
- _derive_responses_function_call_id (fc_ prefix conversion)
- _responses_tools (schema format conversion)
- _chat_messages_to_responses_input (message format conversion)
- _preflight_codex_input_items (input validation)
- _preflight_codex_api_kwargs (API kwargs validation)
- _extract_responses_message_text (response text extraction)
- _extract_responses_reasoning_text (reasoning extraction)
- _normalize_codex_response (full response normalization)
All functions are stateless module-level functions. AIAgent methods remain
as thin one-line wrappers. Both module-level helpers are re-exported from
run_agent.py for backward compatibility with existing test imports.
Includes multimodal inline image support (PR #12969) that the original PR
was missing.
Based on PR #12975 by @kshitijk4poor.
- Fix critical regression: on Linux, Ctrl+C could not interrupt/clear/exit
because isAction(key,'c') shadowed the isCtrl block (both resolve to k.ctrl
on non-macOS). Restructured: isAction block now falls through to interrupt
logic on non-macOS when no selection exists.
- Remove double pbcopy: ink's copySelection() already calls setClipboard()
which handles pbcopy+tmux+OSC52. The extra writeClipboardText call in
useInputHandlers copySelection() was firing pbcopy a second time.
- Remove allowClipboardHotkeys prop from TextInput — every caller passed
isMac, and TextInput already imports isMac. Eliminated prop-drilling
through appLayout, maskedPrompt, and prompts.
- Remove dead code: the isCtrl copy paths (lines 277-288) were unreachable
on any platform after the isAction block changes.
- Simplify textInput Cmd+C: use writeClipboardText directly without the
redundant OSC52 fallback (this path is macOS-only where pbcopy works).
Make the Ink TUI match macOS keyboard expectations: Command handles copy and common editor/session shortcuts, while Control remains reserved for interrupt/cancel flows. Update the visible hotkey help to show platform-appropriate labels.
Follow up salvaged PR #12668 by threading base_url through the
remaining direct-call sites so kimi-k2.5 uses temperature=1.0 on
api.moonshot.ai and keeps 0.6 on api.kimi.com/coding. Add focused
regression tests for run_agent, trajectory_compressor, and
mini_swe_runner.
- only use the native adapter for the canonical Gemini native endpoint
- keep custom and /openai base URLs on the OpenAI-compatible path
- preserve Hermes keepalive transport injection for native Gemini clients
- stabilize streaming tool-call replay across repeated SSE events
- add follow-up tests for base_url precedence, async streaming, and duplicate tool-call chunks
- add a native Gemini adapter over generateContent/streamGenerateContent
- switch the built-in gemini provider off the OpenAI-compatible endpoint
- preserve thought signatures and native functionResponse replay
- route auxiliary Gemini clients through the same adapter
- add focused unit coverage plus native-provider integration checks
Commit 4a9c3565 added a reference to `self.config` in
`_check_compression_model_feasibility()` to pass the user-configured
`auxiliary.compression.context_length` to `get_model_context_length()`.
However, `AIAgent` never stores the loaded config dict as an instance
attribute — the config is loaded into a local variable `_agent_cfg` in
`__init__()` and discarded after init.
This causes an `AttributeError: 'AIAgent' object has no attribute
'config'` on every session start when compression is enabled, caught by
the try/except and logged as a non-fatal DEBUG message.
Fix: store the loaded config as `self._config` in `__init__()` and
update the reference in the feasibility check to use `self._config`.
Follow-up on top of the helix4u #12388 cherry-picks:
- make deferred post-delivery callbacks generation-aware end-to-end so
stale runs cannot clear callbacks registered by a fresher run for the
same session
- bind callback ownership to the active session event at run start and
snapshot that generation inside base adapter processing so later event
mutation cannot retarget cleanup
- pass run_generation through proxy mode and drop stale proxy streams /
final results the same way local runs are dropped
- centralize stop/new interrupt cleanup into one helper and replace the
open-coded branches with shared logic
- unify internal control interrupt reason strings via shared constants
- remove the return from base.py's finally block so cleanup no longer
swallows cancellation/exception flow
- add focused regressions for generation forwarding, proxy stale
suppression, and newer-callback preservation
This addresses all review findings from the initial #12388 review while
keeping the fix scoped to stale-output/typing-loop interrupt handling.
Follow-up on top of the helix4u #6392 cherry-pick:
- reuse one helper for actionable Docker-local file-not-found errors
across document/image/video/audio local-media send paths
- include /outputs/... alongside /output/... in the container-local
path hint
- soften the gateway startup warning so it does not imply custom
host-visible mounts are broken; the warning now targets the specific
risky pattern of emitting container-local MEDIA paths without an
explicit export mount
- add focused regressions for /outputs/... and non-document media hint
coverage
This keeps the salvage aligned with the actual MEDIA delivery problem on
current main while reducing false-positive operator messaging.
Follow-up for the helix4u easy-fix salvage batch:
- route remaining context-engine quiet-mode output through
_should_emit_quiet_tool_messages() so non-CLI/library callers stay
silent consistently
- drop the extra senderAliases computation from WhatsApp allowlist-drop
logging and remove the now-unused import
This keeps the batch scoped to the intended fixes while avoiding
leaked quiet-mode output and unnecessary duplicate work in the bridge.
The cherry-picked commit from #11434 uses the 154585401+ prefixed
noreply format. Add it alongside the existing bare entry so the
contributor audit passes.
New skill: creative/touchdesigner — control a running TouchDesigner
instance via REST API. Build real-time visual networks programmatically.
Architecture:
Hermes Agent -> HTTP REST (curl) -> TD WebServer DAT -> TD Python env
Key features:
- Custom API handler (scripts/custom_api_handler.py) that creates a
self-contained WebServer DAT + callback in TD. More reliable than the
official mcp_webserver_base.tox which frequently fails module imports.
- Discovery-first workflow: never hardcode TD parameter names. Always
probe the running instance first since names change across versions.
- Persistent setup: save the TD project once with the API handler baked
in. TD auto-opens the last project on launch, so port 9981 is live
with zero manual steps after first-time setup.
- Works via curl in execute_code (no MCP dependency required).
- Optional MCP server config for touchdesigner-mcp-server npm package.
Skill structure (2823 lines total):
SKILL.md (209 lines) — setup, workflow, key rules, operator reference
references/pitfalls.md (276 lines) — 24 hard-won lessons
references/operators.md (239 lines) — all 6 operator families
references/network-patterns.md (589 lines) — audio-reactive, generative,
video processing, GLSL, instancing, live performance recipes
references/mcp-tools.md (501 lines) — 13 MCP tool schemas
references/python-api.md (443 lines) — TD Python scripting patterns
references/troubleshooting.md (274 lines) — connection diagnostics
scripts/custom_api_handler.py (140 lines) — REST API handler for TD
scripts/setup.sh (152 lines) — prerequisite checker
Tested on TouchDesigner 099 Non-Commercial (macOS/darwin).
* fix(kimi): force fixed temperature on kimi-k2.* models (k2.5, thinking, turbo)
The prior override only matched the literal model name "kimi-for-coding",
but Moonshot's coding endpoint is hit with real model IDs such as
`kimi-k2.5`, `kimi-k2-turbo-preview`, `kimi-k2-thinking`, etc. Those
requests bypassed the override and kept the caller's temperature, so
Moonshot returns HTTP 400 "invalid temperature: only 0.6 is allowed for
this model" (or 1.0 for thinking variants).
Match the whole kimi-k2.* family:
* kimi-k2-thinking / kimi-k2-thinking-turbo -> 1.0 (thinking mode)
* all other kimi-k2.* -> 0.6 (non-thinking / instant mode)
Also accept an optional vendor prefix (e.g. `moonshotai/kimi-k2.5`) so
aggregator routings are covered.
* refactor(kimi): whitelist-match kimi coding models instead of prefix
Addresses review feedback on PR #12144.
- Replace `startswith("kimi-k2")` with explicit frozensets sourced from
Moonshot's kimi-for-coding model list. The prefix match would have also
clamped `kimi-k2-instruct` / `kimi-k2-instruct-0905`, which are the
separate non-coding K2 family with variable temperature (recommended 0.6
but not enforced — see huggingface.co/moonshotai/Kimi-K2-Instruct).
- Confirmed via platform.kimi.ai docs that all five coding models
(k2.5, k2-turbo-preview, k2-0905-preview, k2-thinking, k2-thinking-turbo)
share the fixed-temperature lock, so the preview-model mapping is no
longer an assumption.
- Drop the fragile `"thinking" in bare` substring test for a set lookup.
- Log a debug line on each override so operators can see when Hermes
silently rewrites temperature.
- Update class docstring. Extend the negative test to parametrize over
kimi-k2-instruct, Kimi-K2-Instruct-0905, and a hypothetical future
kimi-k2-experimental name — all must keep the caller's temperature.
- /retry: use session['history'] instead of non-existent
agent.conversation_history; truncate history at last user message
to match CLI retry_last() behavior; add history_lock safety
- /plan: pass user instruction (arg) to build_plan_path instead of
session_key; add runtime_note so agent knows where to save the plan
- ANSI tool results: render full text via <Ansi wrap=truncate-end>
instead of slicing raw ANSI through compactPreview (which cuts
mid-escape-sequence producing garbled output)
- Move _PENDING_INPUT_COMMANDS frozenset to module level
- Use get_skill_commands() (cached) instead of scan_skill_commands()
(rescans disk) in slash.exec skill interception
- Add 3 retry tests: happy path with history truncation verification,
empty history error, multipart content extraction
- Update test mock target from scan_skill_commands to get_skill_commands
Additional TUI fixes discovered in the same audit:
1. /plan slash command was silently lost — process_command() queues the
plan skill invocation onto _pending_input which nobody reads in the
slash worker subprocess. Now intercepted in slash.exec and routed
through command.dispatch with a new 'send' dispatch type.
Same interception added for /retry, /queue, /steer as safety nets
(these already have correct TUI-local handlers in core.ts, but the
server-side guard prevents regressions if the local handler is
bypassed).
2. Tool results were stripping ANSI escape codes — the messageLine
component used stripAnsi() + plain <Text> for tool role messages,
losing all color/styling from terminal, search_files, etc. Now
uses <Ansi> component (already imported) when ANSI is detected.
3. Terminal tab title now shows model + busy status via useTerminalTitle
hook from @hermes/ink (was never used). Users can identify Hermes
tabs and see at a glance whether the agent is busy or ready.
4. Added 'send' variant to CommandDispatchResponse type + asCommandDispatch
parser + createSlashHandler handler for commands that need to inject
a message into the conversation (plan, queue fallback, steer fallback).
Two TUI fixes:
1. Hyperlinks are now clickable (Cmd+Click / Ctrl+Click) in terminals
that support OSC 8. The markdown renderer was rendering links as
plain colored text — now wraps them in the existing <Link> component
from @hermes/ink which emits OSC 8 escape sequences.
2. Skill slash commands (e.g. /hermes-agent-dev) now work in the TUI.
The slash.exec handler was delegating to the _SlashWorker subprocess
which calls cli.process_command(). For skills, process_command()
queues the invocation message onto _pending_input — a Queue that
nobody reads in the worker subprocess. The skill message was lost.
Now slash.exec detects skill commands early and rejects them so
the TUI falls through to command.dispatch, which correctly builds
and returns the skill payload for the client to send().
Move moonshotai/kimi-k2.5 to position #1 in every model picker list:
- OPENROUTER_MODELS (with 'recommended' tag)
- _PROVIDER_MODELS: nous, kimi-coding, opencode-zen, opencode-go, alibaba, huggingface
- _model_flow_kimi() Coding Plan model list in main.py
kimi-coding-cn and moonshot lists already had kimi-k2.5 first.
Salvage of PR #8018 by @alt-glitch onto current main.
On sandbox teardown, FileSyncManager now downloads the remote .hermes/
directory, diffs against SHA-256 hashes of what was originally pushed,
and applies only changed files back to the host.
Core (tools/environments/file_sync.py):
- sync_back(): orchestrates download -> unpack -> diff -> apply with:
- Retry with exponential backoff (3 attempts, 2s/4s/8s)
- SIGINT trap + defer (prevents partial writes on Ctrl-C)
- fcntl.flock serialization (concurrent gateway sandboxes)
- Last-write-wins conflict resolution with warning
- New remote files pulled back via _infer_host_path prefix matching
Backends:
- SSH: _ssh_bulk_download — tar cf - piped over SSH
- Modal: _modal_bulk_download — exec tar cf - -> proc.stdout.read
- Daytona: _daytona_bulk_download — exec tar cf -> SDK download_file
- All three call sync_back() at the top of cleanup()
Fixes applied during salvage (vs original PR #8018):
| # | Issue | Fix |
|---|-------|-----|
| C1 | import fcntl unconditional — crashes Windows | try/except with fallback; _sync_back_locked skips locking when fcntl=None |
| W1 | assert for runtime guard (stripped by -O) | Replaced with proper if/raise RuntimeError |
| W2 | O(n*m) from _get_files_fn() called per file | Cache mapping once at start of _sync_back_impl, pass to resolve/infer |
| W3 | Dead BulkDownloadFn imports in 3 backends | Removed unused imports |
| W4 | Modal hardcodes root/.hermes, no explanation | Added docstring comment explaining Modal always runs as root |
| S1 | SHA-256 computed for new files where pushed_hash=None | Skip hashing when pushed_hash is None (comparison always False) |
| S2 | Daytona /tmp/.hermes_sync.tar never cleaned up | Added rm -f after download (best-effort) |
Tests: 49 passing (17 new: _infer_host_path edge cases, SIGINT
main/worker thread, Windows fcntl=None fallback, Daytona tar cleanup).
Based on #8018 by @alt-glitch.
Models may send whitespace-only strings like {"conclusion": " "} which
pass bool() but create meaningless conclusions. Strip both inputs so
whitespace-only values are treated as empty.
Adds tests for whitespace-only conclusion and delete_id.
Reviewed-by: @erosika
Add 11 community contributors whose work was cherry-picked via
salvage PRs during the April 16 triage session. Without these
entries, contributor_audit strict mode fails for release attribution.
Contributors: sontianye, jackjin1997, danieldoderlein, lrawnsley,
taeuk178, ogzerber, cola-runner, ygd58, vominh1919, LeonSGP43,
Lubrsy706
Co-authored-by: kshitijk4poor <kshitijk4poor@users.noreply.github.com>
- Extract duplicated activity-callback polling into shared
touch_activity_if_due() helper in tools/environments/base.py
- Use helper from both base.py _wait_for_process and
code_execution_tool.py local polling loop (DRY)
- Add test assertion that timeout output field contains the
timeout message and emoji (#10807)
- Add stream_consumer test for tool-boundary fallback scenario
where continuation is empty but final_text differs from
visible prefix (#10807)
copilot_model_api_mode() called normalize_copilot_model_id() which
fetched the GitHub model catalog via HTTP, then the secondary endpoint
check fetched it again because the catalog was never passed through.
Fix: fetch the catalog once at the top of copilot_model_api_mode()
and pass it to normalize_copilot_model_id(). The secondary check
then sees a non-None catalog and skips the redundant fetch.
For a Claude model switch on Copilot this eliminates one 5-second-
timeout HTTP call from the interactive /model path.
Surfaced during review of PR #10533.
Co-authored-by: kshitijk4poor <kshitijk4poor@users.noreply.github.com>
All 10 call sites in gateway/run.py and gateway/platforms/api_server.py
are inside async functions where a loop is guaranteed to be running.
get_event_loop() is deprecated since Python 3.10 — it can silently
create a new loop when none is running, masking bugs.
get_running_loop() raises RuntimeError instead, which is safer.
Surfaced during review of PRs #10533 and #10647.
Co-authored-by: kshitijk4poor <kshitijk4poor@users.noreply.github.com>
Salvaged from PR #10643 by kshitijk4poor, updated for current main.
Root causes fixed:
1. Telegram xdist mock pollution — new tests/gateway/conftest.py with shared
mock that runs at collection time (prevents ChatType=None caching)
2. VIRTUAL_ENV env var leak — monkeypatch.delenv in _detect_venv_dir tests
3. Copilot base_url missing — add fallback in _resolve_runtime_from_pool_entry
4. Stale vision model assertion — zai now uses glm-5v-turbo
5. Reasoning item id intentionally stripped — assert 'id' not in (store=False)
6. Context length warning unreachable — pass base_url to AIAgent in test
7. Kimi provider label updated — 'Kimi / Kimi Coding Plan' matches models.py
8. Google Workspace calendar tests — rewritten for current production code,
properly mock subprocess on api_module, removed stale +agenda assertions
9. Credential pool auto-seeding — mock _select_pool_entry / _resolve_auto /
_import_codex_cli_tokens to prevent real credentials from leaking into tests
- Populate watcher_* routing fields for watch-only processes (not just
notify_on_complete), so watch-pattern events carry direct metadata
instead of relying solely on session_key parsing fallback
- Extract _parse_session_key() helper to dedupe session key parsing
at two call sites in gateway/run.py
- Add negative test proving cross-thread leakage doesn't happen
- Add edge-case tests for _build_process_event_source returning None
(empty evt, invalid platform, short session_key)
- Add unit tests for _parse_session_key helper
Follow-up to #10459 (salvage of #7527). The copy_context() fix propagates
ALL ContextVars into the cron worker thread, including credential_files.
This test verifies that skill-declared required_credential_files are
visible inside the worker thread, matching the existing env_passthrough
regression test.
PR #9467 added a call to self._fuzzy_file_completions() inside
_context_completions(), but the method was still decorated with
@staticmethod and didn't receive self. Every @ mention in the input
triggers 'name self is not defined' from prompt_toolkit's async
completer, spamming the error on every keystroke.
Fix: remove @staticmethod, add self parameter. The method already uses
self._fuzzy_file_completions() and self._get_project_files() via that
call chain, so it was never meant to stay static after the fuzzy search
feature was added.
- Add GET /api/model/info endpoint that resolves model metadata using the
same 10-step context-length detection chain the agent uses. Returns
auto-detected context length, config override, effective value, and
model capabilities (tools, vision, reasoning, max output, model family).
- Surface model.context_length as model_context_length virtual field in
the config normalize/denormalize cycle. 0 = auto-detect (default),
positive value overrides. Writing 0 removes context_length from the
model dict on disk.
- Add ModelInfoCard component showing resolved context window (e.g. '1M
auto-detected' or '500K override — auto: 1M'), max output tokens, and
colored capability badges (Tools, Vision, Reasoning, model family).
- Inject ModelInfoCard between model field and context_length override in
ConfigPage General tab. Card re-fetches on model change and after save.
- Insert model_context_length right after model in CONFIG_SCHEMA ordering
so the three elements (model input → info card → override) are adjacent.
Add OAuth provider management to the Hermes dashboard with full
lifecycle support for Anthropic (PKCE), Nous and OpenAI Codex
(device-code) flows.
## Backend (hermes_cli/web_server.py)
- 6 new API endpoints:
GET /api/providers/oauth — list providers with connection status
POST /api/providers/oauth/{id}/start — initiate PKCE or device-code
POST /api/providers/oauth/{id}/submit — exchange PKCE auth code
GET /api/providers/oauth/{id}/poll/{session} — poll device-code
DELETE /api/providers/oauth/{id} — disconnect provider
DELETE /api/providers/oauth/sessions/{id} — cancel pending session
- OAuth constants imported from anthropic_adapter (no duplication)
- Blocking I/O wrapped in run_in_executor for async safety
- In-memory session store with 15-minute TTL and automatic GC
- Auth token required on all mutating endpoints
## Frontend
- OAuthLoginModal — PKCE (paste auth code) and device-code (poll) flows
- OAuthProvidersCard — status, token preview, connect/disconnect actions
- Toast fix: createPortal to document.body for correct z-index
- App.tsx: skip animation key bump on initial mount (prevent double-mount)
- Integrated into the Env/Keys page
Cherry-picked from PR #7702 by kshitijk4poor.
Adds Xiaomi MiMo as a direct provider (XIAOMI_API_KEY) with models:
- mimo-v2-pro (1M context), mimo-v2-omni (256K, multimodal), mimo-v2-flash (256K, cheapest)
Standard OpenAI-compatible provider checklist: auth.py, config.py, models.py,
main.py, providers.py, doctor.py, model_normalize.py, model_metadata.py,
models_dev.py, auxiliary_client.py, .env.example, cli-config.yaml.example.
Follow-up: vision tasks use mimo-v2-omni (multimodal) instead of the user's
main model. Non-vision aux uses the user's selected model. Added
_PROVIDER_VISION_MODELS dict for provider-specific vision model overrides.
On failure, falls back to aggregators (gemini flash) via existing fallback chain.
Corrects pre-existing context lengths: mimo-v2-pro 1048576→1000000,
mimo-v2-omni 1048576→256000, adds mimo-v2-flash 256000.
36 tests covering registry, aliases, auto-detect, credentials, models.dev,
normalization, URL mapping, providers module, doctor, aux client, vision
model override, and agent init.
Cherry-picked from PR #7749 by kshitijk4poor with modifications:
- Raise hard image limit from 5 MB to 20 MB (matches most restrictive provider)
- Send images at full resolution first; only auto-resize to 5 MB on API failure
- Add _is_image_size_error() helper to detect size-related API rejections
- Auto-resize uses Pillow (soft dep) with progressive downscale + JPEG quality reduction
- Fix get_model_capabilities() to check modalities.input for vision support
- Increase default vision timeout from 30s to 120s (matches hardcoded fallback intent)
- Applied retry-with-resize to both vision_analyze_tool and browser_vision
Closes#7740
Based on PR #7285 by @kshitijk4poor.
Two bugs affecting Qwen OAuth users:
1. Wrong context window — qwen3-coder-plus showed 128K instead of 1M.
Added specific entries before the generic qwen catch-all:
- qwen3-coder-plus: 1,000,000 (corrected from PR's 1,048,576 per
official Alibaba Cloud docs and OpenRouter)
- qwen3-coder: 262,144
2. Random stopping — max_tokens was suppressed for Qwen Portal, so the
server applied its own low default. Reasoning models exhaust that on
thinking tokens. Now: honor explicit max_tokens, default to 65536
when unset.
Co-authored-by: kshitijk4poor <82637225+kshitijk4poor@users.noreply.github.com>
async_call_llm (and call_llm) can return non-OpenAI objects from
custom providers or adapter shims, crashing downstream consumers
with misleading AttributeError ('str' has no attribute 'choices').
Add _validate_llm_response() that checks the response has the
expected .choices[0].message shape before returning. Wraps all
return paths in call_llm, async_call_llm, and fallback paths.
Fails fast with a clear RuntimeError identifying the task, response
type, and a preview of the malformed payload.
Closes#7264
Four fixes to auxiliary_client.py:
1. Respect explicit provider as hard constraint (#7559)
When auxiliary.{task}.provider is explicitly set (not 'auto'),
connection/payment errors no longer silently fallback to cloud
providers. Local-only users (Ollama, vLLM) will no longer get
unexpected OpenRouter billing from auxiliary tasks.
2. Eliminate model='default' sentinel (#7512)
_resolve_api_key_provider() no longer sends literal 'default' as
model name to APIs. Providers without a known aux model in
_API_KEY_PROVIDER_AUX_MODELS are skipped instead of producing
model_not_supported errors.
3. Add payment/connection fallback to async_call_llm (#7512)
async_call_llm now mirrors sync call_llm's fallback logic for
payment (402) and connection errors. Previously, async consumers
(session_search, web_tools, vision) got hard failures with no
recovery. Also fixes hardcoded 'openrouter' fallback to use the
full auto-detection chain.
4. Use accurate error reason in fallback logs (#7512)
_try_payment_fallback() now accepts a reason parameter and uses
it in log messages. Connection timeouts are no longer misleadingly
logged as 'payment error'.
Closes#7559Closes#7512
The auxiliary client always calls client.chat.completions.create(),
ignoring the api_mode config flag. This breaks codex-family models
(e.g. gpt-5.3-codex) on direct OpenAI API keys, which need the
/v1/responses endpoint.
Changes:
- Expand _resolve_task_provider_model to return api_mode (5-tuple)
- Read api_mode from auxiliary.{task}.api_mode config and env vars
(AUXILIARY_{TASK}_API_MODE)
- Pass api_mode through _get_cached_client to resolve_provider_client
- Add _needs_codex_wrap/_wrap_if_needed helpers that wrap plain OpenAI
clients in CodexAuxiliaryClient when api_mode=codex_responses or
when auto-detection finds api.openai.com + codex model pattern
- Apply wrapping at all custom endpoint, named custom provider, and
API-key provider return paths
- Update test mocks for the new 5-tuple return format
Users can now set:
auxiliary:
compression:
model: gpt-5.3-codex
base_url: https://api.openai.com/v1
api_mode: codex_responses
Closes#6800
Aligns MiniMax provider with official API documentation. Fixes 6 bugs:
transport mismatch (openai_chat -> anthropic_messages), credential leak
in switch_model(), prompt caching sent to non-Anthropic endpoints,
dot-to-hyphen model name corruption, trajectory compressor URL routing,
and stale doctor health check.
Also corrects context window (204,800), thinking support (manual mode),
max output (131,072), and model catalog (M2 family only on /anthropic).
Source: https://platform.minimax.io/docs/api-reference/text-anthropic-api
Co-authored-by: kshitijk4poor <kshitijk4poor@users.noreply.github.com>
When the model calls terminal() in foreground mode without background=true
(e.g. to start a server), the tool call blocks until the command exits or
the timeout expires. Without an upper bound the model can request arbitrarily
high timeouts (the schema had minimum=1 but no maximum), blocking the entire
agent session for hours until the gateway idle watchdog kills it.
Changes:
- Add FOREGROUND_MAX_TIMEOUT (600s, configurable via
TERMINAL_MAX_FOREGROUND_TIMEOUT env var) that caps foreground timeout
- Clamp effective_timeout to the cap when background=false and timeout
exceeds the limit
- Include a timeout_note in the tool result when clamped, nudging the
model to use background=true for long-running processes
- Update schema description to show the max timeout value
- Remove dead clamping code in the background branch that could never
fire (max_timeout was set to effective_timeout, so timeout > max_timeout
was always false)
- Add 7 tests covering clamping, no-clamping, config-default-exceeds-cap
edge case, background bypass, default timeout, constant value, and
schema content
Self-review fixes:
- Fixed bug where timeout_note said 'Requested timeout Nones' when
clamping fired from config default exceeding cap (timeout param is
None). Now uses unclamped_timeout instead of the raw timeout param.
- Removed unused pytest import from test file
- Extracted test config dict into _make_env_config() helper
- Fixed tautological test_default_value assertion
- Added missing test for config default > cap with no model timeout
The gateway /model command stored session overrides in
_session_model_overrides but run_sync() never consulted them when
resolving the model and runtime for the next message. It always read
from config.yaml, so the switch was lost as soon as a new agent was
created.
Two fixes:
1. In run_sync(), apply _session_model_overrides after resolving from
config.yaml/env — the override takes precedence for model, provider,
api_key, base_url, and api_mode.
2. In post-run fallback detection, check whether the model mismatch
(agent.model != config_model) is due to an intentional /model switch
before evicting the cached agent. Without this, the first message
after /model would work (cached agent reused) but the fallback
detector would evict it, causing the next message to revert.
Affects all gateway platforms (Telegram, Discord, Slack, WhatsApp,
Signal, Matrix, BlueBubbles, HomeAssistant) since they all share
GatewayRunner._run_agent().
Fixes#6213
kshitijk4poor