…(optional via config)

- Translate severity levels (HIGH→高危, CRITICAL→致命, etc.)
- Translate common rule titles (Pipe to interpreter→管道传输到解释器, etc.)
- Translate 'Security scan' prefix to '安全扫描'
- Add rule_id-based lookup for comprehensive rule coverage
- Fallback to title string matching for rules not in the rule map
- Respects approvals.language config (zh/en)

Add a circuit breaker mechanism in _invoke_tool() and the sequential
execution path to detect and break infinite tool-call loops.

When the same tool with identical arguments is called 3 consecutive
times, the circuit breaker triggers and returns an error, forcing the
LLM to try a different strategy.

The breaker uses MD5 hashing of (tool_name, json_args) to create a
stable signature, then checks if the last N calls all have the same
signature. After triggering, the counter resets so future calls can
succeed.

- Added _consecutive_tool_calls and _circuit_breaker_threshold to AIAgent.__init__
- Added circuit breaker logic to _invoke_tool() (concurrent path)
- Added circuit breaker logic to _execute_tool_calls_sequential() (sequential path)
- All 22 existing tests pass

…shold

- Change error message from Chinese to English for upstream compatibility
- Increase default threshold from 3 to 5 to avoid false positives
  (legitimate retry scenarios exist, e.g. network jitter)
- Fix indentation bug in sequential execution path
- All 22 existing tests pass

…elp text

- Add _DESCRIPTION_ZH and _CATEGORY_ZH dicts in commands.py
- Add _desc_zh() and _cat_zh() helpers for runtime localization
- Localize COMMANDS, COMMANDS_BY_CATEGORY, gateway_help_lines(), telegram_bot_commands()
- Add HELP_ZH and HOTKEYS_ZH constants in core.ts for web UI
- Add BRAND_ZH and getBrand() in theme.ts for branding strings
- Language detection from config.yaml (approvals.language or language key)
- Backward compatible: English defaults when language != 'zh'

- banner.py: Chinese translations for welcome banner text (tools, skills,
  MCP servers, update messages)
- skin_engine.py: new 'chinese' skin with gold theme, auto-select based
  on approvals.language config
- tips.py: full Chinese tips corpus (TIPS_ZH) with language detection

… split

Prevents the LLM from retrying different variations of a failing tool call.
When a tool returns errors/empty results N times consecutively (regardless of
argument changes), the counter triggers and forces a strategy change.

- Added _tool_failure_count dict and _tool_failure_threshold (5) in __init__
- Added _check_tool_failure() method to detect failures and update counter
- Added check calls in both concurrent and sequential execution paths
- Each tool has independent counter; success resets counter to 0

- Remove erroneous _check_tool_failure call in _invoke_tool that referenced
  undefined variable function_result
- Add temperature field to auxiliary.vision config (default: 0.1)
- Fixes 8 failing tests in concurrent tool execution and camofox vision

- Move _check_tool_failure calls to AFTER each tool execution
- Adds failure retry counter check for all tool types (todo, session_search,
  memory, memory manager tools, clarify, delegate_task, and generic tools)
- Prevents LLM from retrying different parameter variations of failing tools
- Complements the circuit breaker which detects identical tool+args loops

- Add read_timestamps to _read_tracker initialization (file_tools.py)
- Fix camofox vision test mock path (test_browser_camofox.py)
- Catch ModuleNotFoundError in browser_tool camofox import
- Fix _CAMEL_ALIASES apiKey mapping in config.py
- Use os.path.join for hermes_home/.env path (file_safety.py)

Fixes 8 failing tests across 6 test files.

- Fix mock paths in test_zombie_process_cleanup.py (use run_agent.* instead of tools.*)
- Fix mock path in test_agent_cache.py (mock run_agent.cleanup_vm instead of _tt.cleanup_vm)
- Use get_hermes_home() in build_write_denied_paths for profile-aware path resolution
- Add ModuleNotFoundError handling in browser_tool camofox and auxiliary_client imports

Fixes 6 failing tests:
- test_zombie_process_cleanup.py: test_close_calls_cleanup_functions, test_close_survives_partial_failures
- test_agent_cache.py: test_close_vs_release_full_teardown_difference
- test_write_deny.py: test_hermes_env
- test_approval_heartbeat.py: test_wait_returns_immediately_on_user_response, test_heartbeat_import_failure_does_not_break_wait

- Fix build_write_denied_paths to use home/.hermes/.env instead of get_hermes_home()
- Fix mock paths in test_zombie_process_cleanup.py (use tools.process_registry.process_registry)

Fixes 6 failing tests across 4 test files.

pytest's caplog fixture doesn't automatically capture warnings from
hermes_cli.config logger because the logger's level is not set.

This conftest.py sets the logger level to DEBUG for all tests in
tests/hermes_cli/, ensuring caplog captures WARNING and above messages.

Fixes 2 failing tests in test_provider_config_validation.py:

When a tool call fails repeatedly, the compression model is asked
for a suggestion. This gives the main model a fresh perspective
to break out of tool loops.

- Added _get_tool_suggestion() for retry threshold failures
- Added _get_consecutive_suggestion() for identical-call circuit breaker
- Suggestions are appended to the error message with a 💡 emoji
- Gracefully falls back if compression model is unavailable

When a tool call fails repeatedly, suggestions are provided even
without a compression model configured.

- Added _heuristic_suggestion() — pattern matching on error messages
- Added _heuristic_consecutive_suggestion() — 20+ tool-specific suggestions
- Compression model is tried first, then falls back to heuristics
- Works zero-config — no LLM needed for basic suggestions

Replaced complex pattern-matching heuristics with simple generic prompts.
- Removed _heuristic_suggestion() (8 pattern rules)
- Removed _heuristic_consecutive_suggestion() (20+ tool mappings)
- Fallback is now just a direct text suggestion
- Compression model still tries first for smarter suggestions
- Net: -90 lines of dead pattern logic

Added README explaining:
- This is a fork of NousResearch/hermes-agent
- What Hermes Agent is
- Custom modifications in this fork (circuit breaker improvements, test fixes)

…role

The integration tests call write_file_tool which goes through
_get_file_ops() that creates a terminal environment. Without
explicit TERMINAL_ENV=local, the CI environment may default to
modal which fails due to missing credentials.

Fixes: test_net_new_file_no_warning, test_sibling_agent_write_surfaces_warning_through_handler

The test_terminal_and_file_toolsets_resolve_all_tools and
test_terminal_tool_present tests call get_tool_definitions() which
filters tools based on check_terminal_requirements(). Without
explicit TERMINAL_ENV=local, the CI environment's TERMINAL_ENV=modal
(without credentials) causes check_terminal_requirements() to return
False, dropping the terminal tool from the schema.

Fixes: test_terminal_and_file_toolsets_resolve_all_tools,
       test_terminal_tool_present

…erflow