mirrors/hermes-agent
1
0
Fork 0
mirror of https://github.com/NousResearch/hermes-agent.git synced 2026-04-25 00:51:20 +00:00
Code Issues Projects Releases Packages Wiki Activity Actions
hermes-agent/tests
History
Teknium b2b4a9ee7d
fix(gateway): hygiene compression ignores config context_length and 1.4x exceeds model limit 
Three bugs in gateway session hygiene pre-compression caused 'Session too
large' errors for ~200K context models like GLM-5-turbo on z.ai:

1. Gateway hygiene called get_model_context_length(model) without passing
   config_context_length, provider, or base_url — so user overrides like
   model.context_length: 180000 were ignored, and provider-aware detection
   (models.dev, z.ai endpoint) couldn't fire. The agent's own compressor
   correctly passed all three (run_agent.py line 1038).

2. The 1.4x safety factor on rough token estimates pushed the compression
   threshold above the model's actual context limit:
     200K * 0.85 * 1.4 = 238K > 200K (model limit)
   So hygiene never compressed, sessions grew past the limit, and the API
   rejected the request.

3. Same issue for the warn threshold: 200K * 0.95 * 1.4 = 266K.

Fix:
- Read model.context_length, provider, and base_url from config.yaml
  (same as run_agent.py does) and pass them to get_model_context_length()
- Resolve provider/base_url from runtime when not in config
- Cap the 1.4x-adjusted compress threshold at 95% of context_length
- Cap the 1.4x-adjusted warn threshold at context_length

Affects: z.ai GLM-5/GLM-5-turbo, any ~200K or smaller context model
where the 1.4x factor would push 85% above 100%.

Ref: Discord report from Ddox — glm-5-turbo on z.ai coding plan
2026-03-22 15:15:37 -07:00
..
acp fix(acp): preserve session provider when switching models 2026-03-21 15:54:10 -07:00
agent fix(tests): resolve all consistently failing tests 2026-03-22 05:58:26 -07:00
cron fix(cron): support Telegram topic delivery via platform:chat_id:thread_id format (#2455) 2026-03-22 04:18:28 -07:00
fakes fix: streaming tool call parsing, error handling, and fake HA state mutation 2026-03-14 14:27:20 +03:00
gateway fix(gateway): hygiene compression ignores config context_length and 1.4x exceeds model limit 2026-03-22 15:15:37 -07:00
hermes_cli Merge pull request #2465 from NousResearch/hermes/hermes-31d7db3b 2026-03-22 04:56:48 -07:00
honcho_integration feat(honcho): instance-local config via HERMES_HOME, default session strategy to per-directory 2026-03-21 09:34:00 -07:00
integration feat(web): add Parallel as alternative web search/extract backend (#1696) 2026-03-17 04:02:02 -07:00
skills fix: persist google oauth pkce for headless auth 2026-03-14 22:11:34 -07:00
tools fix(mcp-oauth): port mismatch, path traversal, and shared handler state (salvage #2521) (#2552) 2026-03-22 15:02:26 -07:00
__init__.py A bit of restructuring for simplicity and organization 2025-10-01 23:29:25 +00:00
conftest.py fix(approval): show full command in dangerous command approval (#1553) 2026-03-17 02:02:33 -07:00
run_interrupt_test.py fix: thread safety for concurrent subagent delegation (#1672) 2026-03-17 02:53:33 -07:00
test_413_compression.py feat: improve context compaction handoff summaries (#1273) 2026-03-14 02:33:31 -07:00
test_860_dedup.py fix: eliminate 3x SQLite message duplication in gateway sessions (#860) 2026-03-10 15:22:44 -07:00
test_1630_context_overflow_loop.py fix: prevent infinite 400 loop on context overflow + block prompt injection via cache files (#1630, #1558) 2026-03-17 01:50:59 -07:00
test_agent_guardrails.py feat: pre-call sanitization and post-call tool guardrails (#1732) 2026-03-17 04:24:27 -07:00
test_agent_loop.py fix: salvage gateway dedup and executor cleanup from PR #993 2026-03-14 11:03:20 -07:00
test_agent_loop_tool_calling.py fix: skip hanging tests + add global test timeout 2026-03-12 01:23:28 -07:00
test_agent_loop_vllm.py test: restore vllm integration coverage and add dict-args regression 2026-03-15 08:02:29 -07:00
test_anthropic_adapter.py fix(prompt-caching): skip top-level cache_control on role:tool for OpenRouter 2026-03-21 16:54:43 -07:00
test_anthropic_error_handling.py fix(anthropic): retry 429/529 errors and surface error details to users 2026-03-17 01:07:11 +03:00
test_anthropic_oauth_flow.py fix: preflight Anthropic auth and prefer Claude store 2026-03-14 19:38:55 -07:00
test_anthropic_provider_persistence.py fix: preflight Anthropic auth and prefer Claude store 2026-03-14 19:38:55 -07:00
test_api_key_providers.py fix: resolve MiniMax 401 auth error by defaulting to anthropic_messages (#2103) 2026-03-19 17:47:05 -07:00
test_atomic_json_write.py test: cover atomic temp cleanup on interrupts 2026-03-14 22:31:51 -07:00
test_atomic_yaml_write.py test: cover atomic temp cleanup on interrupts 2026-03-14 22:31:51 -07:00
test_auth_codex_provider.py refactor(auth): transition Codex OAuth tokens to Hermes auth store 2026-03-01 19:59:24 -08:00
test_auth_nous_provider.py Fix nous refresh token rotation failure in case where api key mint/retrieval fails 2026-03-02 17:18:15 +11:00
test_auxiliary_config_bridge.py feat(compression): add summary_base_url + move compression config to YAML-only 2026-03-17 04:46:15 -07:00
test_batch_runner_checkpoint.py fix: sanitize chat payloads and provider precedence 2026-03-13 23:59:12 -07:00
test_cli_approval_ui.py fix(cli): repair dangerous command approval UI 2026-03-14 11:57:44 -07:00
test_cli_extension_hooks.py refactor(cli): add protected TUI extension hooks for wrapper CLIs 2026-03-21 09:42:07 -07:00
test_cli_init.py fix: skip model auto-detection for custom/local providers 2026-03-20 04:35:17 -07:00
test_cli_interrupt_subagent.py fix: thread safety for concurrent subagent delegation (#1672) 2026-03-17 02:53:33 -07:00
test_cli_loading_indicator.py fix(cli): add loading indicators for slow slash commands 2026-03-10 17:31:00 -07:00
test_cli_mcp_config_watch.py fix: auto-reload MCP tools when mcp_servers config changes without restart (#1474) 2026-03-15 19:03:34 -07:00
test_cli_model_command.py feat: auto-detect provider when switching models via /model (#1506) 2026-03-16 04:34:45 -07:00
test_cli_new_session.py fix: complete session reset — missing compressor counters + test 2026-03-20 04:35:17 -07:00
test_cli_plan_command.py fix: save /plan output in workspace (#1381) 2026-03-14 21:28:51 -07:00
test_cli_prefix_matching.py feat: add /tools disable/enable/list slash commands with session reset (#1652) 2026-03-17 02:05:26 -07:00
test_cli_preloaded_skills.py feat: preload CLI skills on launch (#1359) 2026-03-14 19:33:59 -07:00
test_cli_provider_resolution.py feat: overhaul context length detection with models.dev and provider-aware resolution (#2158) 2026-03-20 06:04:33 -07:00
test_cli_retry.py test: lock retry replacement semantics 2026-03-14 21:19:22 -07:00
test_cli_secret_capture.py feat: secure skill env setup on load (core #688) 2026-03-13 03:14:04 -07:00
test_cli_skin_integration.py fix(test): add missing voice state attrs to CLI stub in skin tests 2026-03-14 15:00:45 +03:00
test_cli_status_bar.py feat: add route-aware pricing estimates (#1695) 2026-03-17 03:44:44 -07:00
test_cli_tools_command.py feat: add /tools disable/enable/list slash commands with session reset (#1652) 2026-03-17 02:05:26 -07:00
test_codex_execution_paths.py feat: simple fallback model for provider resilience 2026-03-08 20:22:33 -07:00
test_codex_models.py fix: add codex forward-compat model listing 2026-03-13 21:34:01 -07:00
test_compression_boundary.py fix(agent): prevent silent tool result loss during context compression (#1993) 2026-03-18 15:22:51 -07:00
test_context_pressure.py feat: context pressure warnings for CLI and gateway (#2159) 2026-03-20 08:37:36 -07:00
test_context_references.py feat: @ context references — inline file, folder, diff, git, and URL injection 2026-03-21 15:57:13 -07:00
test_context_token_tracking.py fix(tests): resolve all consistently failing tests 2026-03-22 05:58:26 -07:00
test_dict_tool_call_args.py test: restore vllm integration coverage and add dict-args regression 2026-03-15 08:02:29 -07:00
test_display.py fix: add upstream guard for non-dict function_args + tests for build_tool_preview 2026-03-09 21:01:40 -07:00
test_evidence_store.py feat: add OSS Security Forensics skill (Skills Hub) (#1482) 2026-03-15 21:59:53 -07:00
test_external_credential_detection.py refactor(auth): transition Codex OAuth tokens to Hermes auth store 2026-03-01 19:59:24 -08:00
test_fallback_model.py feat: upgrade MiniMax default to M2.7 + add new OpenRouter models 2026-03-18 02:42:58 -07:00
test_file_permissions.py security: enforce 0600/0700 file permissions on sensitive files (inspired by openclaw) 2026-03-09 02:19:32 -07:00
test_flush_memories_codex.py fix: update all test mocks for call_llm migration 2026-03-11 21:06:54 -07:00
test_hermes_state.py fix: search all sources by default in session_search (#1892) 2026-03-18 02:21:29 -07:00
test_honcho_client_config.py fix(honcho): auto-enable when API key is present 2026-03-01 03:12:37 -05:00
test_insights.py feat: add route-aware pricing estimates (#1695) 2026-03-17 03:44:44 -07:00
test_interactive_interrupt.py fix: thread safety for concurrent subagent delegation (#1672) 2026-03-17 02:53:33 -07:00
test_interrupt_propagation.py fix: thread safety for concurrent subagent delegation (#1672) 2026-03-17 02:53:33 -07:00
test_managed_server_tool_support.py test: fix stale CI assumptions in parser and quick-command coverage (#1236) 2026-03-13 21:56:12 -07:00
test_minisweagent_path.py fix: worktree-aware minisweagent path discovery + clean up requirements check (#1248) 2026-03-13 23:39:51 -07:00
test_model_metadata_local_ctx.py fix: prefer loaded instance context size over max for LM Studio 2026-03-19 21:24:53 +01:00
test_model_provider_persistence.py feat: integrate GitHub Copilot providers across Hermes 2026-03-17 23:40:22 -07:00
test_model_tools.py test: strengthen assertions across 3 more test files (batch 2) 2026-03-05 18:46:30 -08:00
test_model_tools_async_bridge.py fix: use per-thread persistent event loops in worker threads 2026-03-20 15:41:06 -04:00
test_openai_client_lifecycle.py fix: audit fixes — 5 bugs found and resolved 2026-03-16 06:35:46 -07:00
test_personality_none.py feat(cli,gateway): add /personality none and custom personality support 2026-03-09 17:31:54 +03:00
test_plugins.py fix(tests): resolve all consistently failing tests 2026-03-22 05:58:26 -07:00
test_plugins_cmd.py feat(cli): add hermes plugins install/remove/list command 2026-03-21 09:47:33 -07:00
test_provider_parity.py feat: add Vercel AI Gateway provider (#1628) 2026-03-17 00:12:16 -07:00
test_quick_commands.py fix: thread safety for concurrent subagent delegation (#1672) 2026-03-17 02:53:33 -07:00
test_real_interrupt_subagent.py fix: thread safety for concurrent subagent delegation (#1672) 2026-03-17 02:53:33 -07:00
test_reasoning_command.py fix: /reasoning command — add gateway support, fix display, persist settings (#1031) 2026-03-12 05:38:19 -07:00
test_redirect_stdout_issue.py fix: use session_key instead of chat_id for adapter interrupt lookups 2026-03-12 08:35:45 -07:00
test_resume_display.py feat: display previous messages when resuming a session in CLI 2026-03-08 17:45:45 -07:00
test_run_agent.py fix: prevent Anthropic token leaking to third-party anthropic_messages providers (salvage #2383) (#2389) 2026-03-21 16:42:46 -07:00
test_run_agent_codex_responses.py fix(codex): handle reasoning-only responses and replay path (#2070) 2026-03-19 10:34:44 -07:00
test_runtime_provider_resolution.py fix: respect DashScope v1 runtime mode for alibaba (#2459) 2026-03-22 04:24:43 -07:00
test_setup_model_selection.py fix(setup): remove dead code causing is_coding_plan NameError crash 2026-03-13 04:42:26 +03:00
test_sql_injection.py fix(security): eliminate SQL string formatting in execute() calls 2026-03-19 15:16:35 +01:00
test_streaming.py fix: always fall back to non-streaming on ANY streaming error 2026-03-16 06:15:09 -07:00
test_timezone.py fix: skip stale cron jobs on gateway restart instead of firing immediately 2026-03-16 23:48:14 -07:00
test_tool_call_parsers.py fix(mistral-parser): handle nested JSON in fallback extraction 2026-03-21 09:41:17 -07:00
test_toolset_distributions.py test: add unit tests for 8 modules (batch 2) 2026-02-26 13:54:20 +03:00
test_toolsets.py fix: add missing Platform.SIGNAL to toolset mappings, update test + config docs 2026-03-09 23:27:19 -07:00
test_trajectory_compressor.py fix: harden trajectory compressor summary content handling 2026-03-14 11:03:25 -07:00
test_worktree.py fix: harden salvaged worktree include checks 2026-03-14 21:51:27 -07:00
test_worktree_security.py fix: harden salvaged worktree include checks 2026-03-14 21:51:27 -07:00