hermes-agent/agent
Teknium 1a3ae6ac6e
feat: structured API error classification for smart failover (#6514)
Add agent/error_classifier.py with a priority-ordered classification
pipeline that replaces scattered inline string-matching in the retry
loop with structured error taxonomy and recovery hints.

FailoverReason enum (14 categories): auth, auth_permanent, billing,
rate_limit, overloaded, server_error, timeout, context_overflow,
payload_too_large, model_not_found, format_error, thinking_signature,
long_context_tier, unknown.

ClassifiedError dataclass carries reason + recovery action hints
(retryable, should_compress, should_rotate_credential, should_fallback).

Key improvements over inline matching:
- 402 disambiguation: 'insufficient credits' = billing (immediate rotate),
  'usage limit, try again' = rate_limit (backoff first)
- OpenRouter 403 'key limit exceeded' correctly classified as billing
- Error cause chain walking (walks __cause__/__context__ up to 5 levels)
- Body message included in pattern matching (SDK str() misses it)
- Server disconnect + large session check ordered before generic transport
  catch so RemoteProtocolError triggers compression when appropriate
- Chinese error message support for context overflow

run_agent.py: replaced 6 inline detection blocks with classifier calls,
net -55 lines. All recovery actions (pool rotation, fallback activation,
compression, transport recovery) unchanged.

65 new unit tests + 10 E2E tests + live tests with real SDK error objects.
Inspired by OpenClaw's failover error classification system.
2026-04-09 04:10:11 -07:00
..
__init__.py Refactor Terminal and AIAgent cleanup 2026-02-21 22:31:43 -08:00
anthropic_adapter.py fix: normalize httpx.URL base_url + strip thinking signatures for third-party endpoints 2026-04-08 16:39:29 -07:00
auxiliary_client.py fix: restore codex fallback auth-store lookup 2026-04-09 01:56:10 -07:00
builtin_memory_provider.py refactor: add tool_error/tool_result helpers + read_raw_config, migrate 129 callsites 2026-04-07 13:36:38 -07:00
context_compressor.py fix(compaction): token-budget primary tail protection 2026-04-08 23:54:23 -07:00
context_references.py refactor: replace inline HERMES_HOME re-implementations with get_hermes_home() 2026-04-07 10:40:34 -07:00
copilot_acp_client.py fix: bridge tool-calls in copilot-acp adapter 2026-04-06 01:47:57 -07:00
credential_pool.py fix: reduce credential exhaustion TTL from 24 hours to 1 hour (#6504) 2026-04-09 02:37:23 -07:00
display.py refactor: remove 24 confirmed dead functions — 432 lines of unused code 2026-04-07 11:41:26 -07:00
error_classifier.py feat: structured API error classification for smart failover (#6514) 2026-04-09 04:10:11 -07:00
insights.py fix(insights): show cache tokens in overview so total adds up (#4428) 2026-04-01 03:06:47 -07:00
memory_manager.py refactor: add tool_error/tool_result helpers + read_raw_config, migrate 129 callsites 2026-04-07 13:36:38 -07:00
memory_provider.py refactor: codebase-wide lint cleanup — unused imports, dead code, and inefficient patterns (#5821) 2026-04-07 10:25:31 -07:00
model_metadata.py fix: resolve opencode.ai context window to 1M and clean up display formatting 2026-04-09 01:43:22 -07:00
models_dev.py feat(qwen): add Qwen OAuth provider with portal request support 2026-04-08 13:46:30 -07:00
prompt_builder.py feat(gateway): add BlueBubbles iMessage platform adapter (#6437) 2026-04-08 23:54:03 -07:00
prompt_caching.py fix(prompt-caching): skip top-level cache_control on role:tool for OpenRouter 2026-03-21 16:54:43 -07:00
rate_limit_tracker.py feat: capture provider rate limit headers and show in /usage (#6541) 2026-04-09 03:43:14 -07:00
redact.py fix: mem0 API v2 compat, prefetch context fencing, secret redaction (#5423) 2026-04-05 22:43:33 -07:00
retry_utils.py feat(agent): add jittered retry backoff 2026-04-08 00:41:36 -07:00
skill_commands.py feat(skills): add skill config interface + llm-wiki skill (#5635) 2026-04-06 13:49:13 -07:00
skill_utils.py refactor: codebase-wide lint cleanup — unused imports, dead code, and inefficient patterns (#5821) 2026-04-07 10:25:31 -07:00
smart_model_routing.py Merge branch 'main' into rewbs/tool-use-charge-to-subscription 2026-04-02 11:00:35 +11:00
subdirectory_hints.py fix(agent): catch PermissionError in subdirectory hint discovery 2026-04-09 03:10:30 -07:00
title_generator.py feat(agent): configurable timeouts for auxiliary LLM calls via config.yaml (#3597) 2026-03-28 14:35:28 -07:00
trajectory.py Refactor Terminal and AIAgent cleanup 2026-02-21 22:31:43 -08:00
usage_pricing.py fix: status bar shows 26K instead of 260K for token counts with trailing zeros (#3024) 2026-03-25 12:45:58 -07:00