hermes-agent/tests
Teknium d3c167b644
fix(profiles): cross-profile soft guard on file-write tools + system-prompt hint (#31290)
* fix(profiles): cross-profile soft guard on file-write tools + system-prompt hint

Adds a soft guard so an agent running under one Hermes profile cannot
silently edit a different profile's skills/plugins/cron/memories.
Three layers:

A. agent/file_safety.classify_cross_profile_target
   Classifies a write target against the active HERMES_HOME. Returns
   a {active_profile, target_profile, area, target_path} dict when the
   path lands in another profile's scoped area. PROFILE_SCOPED_AREAS =
   (skills, plugins, cron, memories). get_cross_profile_warning()
   wraps it into a model-facing error string that names both profiles,
   names the area, and points at the cross_profile=True bypass.

   Defense-in-depth, NOT a security boundary — the terminal tool runs
   as the same OS user and can write any of these paths directly. The
   guard exists to prevent confused-agent corruption, not to stop a
   determined attacker. SECURITY.md §3.2 (terminal-bypass posture)
   still applies.

   Wired into tools/file_tools.write_file_tool and patch_tool with a
   cross_profile=False kwarg. WRITE_FILE_SCHEMA and PATCH_SCHEMA both
   advertise cross_profile so the model can pass it after explicit
   user direction. patch_tool extracts target paths from V4A patch
   bodies before checking (same shape as the existing sensitive-path
   check).

   skill_manage is already scoped to the active profile's SKILLS_DIR
   by construction, so no extra guard wiring is needed there. The
   D-side error message (below) still names other profiles when the
   skill exists elsewhere.

B. agent/system_prompt
   One deterministic line near the environment-hints block names the
   active profile and tells the model not to modify another profile's
   skills/plugins/cron/memories without explicit direction. Profile
   name is stable for the lifetime of the AIAgent, so the line is
   prompt-cache-safe.

D. tools/skill_manager_tool._skill_not_found_error
   Replaces the bare "Skill 'X' not found." with a message that:
     - names the active profile,
     - searches OTHER profiles' skills dirs for the same name,
     - names the profile(s) where the skill exists and the path,
     - suggests `hermes -p <name>` to switch profiles, or
       cross_profile=True for an explicit edit.

   All 5 "not found" sites in skill_manager_tool (edit, patch, delete,
   write_file, remove_file) now go through the helper.

Reference incident (May 2026): a hermes-security profile session
edited skills under both ~/.hermes/profiles/hermes-security/skills/
AND ~/.hermes/skills/ (the default profile's skills) without
realizing the second path belonged to a different profile. Three of
the four skill files needed manual restoration afterward.

What this PR does NOT do:

  * No hard block. The terminal tool can still touch any of these
    paths with no guard — same posture as the dangerous-command
    approval flow. SECURITY.md §3.2 applies.
  * No regex sweep on terminal commands for cross-profile paths.
    That direction is a Skills-Guard-style arms race (cd + relative
    paths, base64, etc.) and would false-positive on legitimate
    cross-profile reads. Filed as a follow-up.
  * No on-disk path migration. ~/.hermes/skills/ remains the
    default profile's skills dir; this PR is about telling the
    agent about that boundary, not changing the layout.

Tests:
  tests/agent/test_file_safety_cross_profile.py (16 tests)
    - _resolve_active_profile_name covers default/named/failure paths
    - classify_cross_profile_target covers all four scoped areas,
      both directions (default → named, named → default, named → named),
      non-Hermes paths, and root-level config files
    - get_cross_profile_warning covers in-profile no-op, cross-profile
      message shape, and the defense-in-depth self-documentation

  tests/tools/test_cross_profile_guard.py (12 tests)
    - write_file: in-profile allow, cross-profile block, cross_profile=True
      bypass, non-Hermes pass-through
    - patch: replace-mode block, cross_profile=True bypass, V4A patch
      path extraction
    - skill_manage: error names the other profile (single + multiple),
      missing-everywhere falls back to skills_list hint
    - system prompt: contract-level checks (both branches present,
      cross_profile=True mentioned, ~/.hermes/profiles/ referenced)

All 207 existing tests in file_safety/file_operations/skill_manager
still pass. 10 system-prompt tests still pass.

E2E verified: the exact incident scenario (security profile editing
default's hermes-agent-dev skill) is now blocked with the warning
message; cross_profile=True unblocks.

* fix(code_execution): add cross_profile to write_file/patch stubs

The cross_profile kwarg added to write_file_tool/patch_tool needs to
flow through the execute_code sandbox stubs in _TOOL_STUBS so the
test_stubs_cover_all_schema_params drift test passes. Without this,
scripts running inside execute_code couldn't pass cross_profile=True
through hermes_tools.write_file().

Caught by CI on PR #31290.
2026-05-24 00:38:17 -07:00
..
acp test(acp): pin parse_model_input in slash-command tests 2026-05-23 01:44:56 -07:00
acp_adapter feat(azure-foundry): add Microsoft Entra ID auth 2026-05-18 10:14:38 -07:00
agent fix(profiles): cross-profile soft guard on file-write tools + system-prompt hint (#31290) 2026-05-24 00:38:17 -07:00
cli fix(cli): reuse canonical root model key normalization in load_cli_config 2026-05-23 23:08:05 -07:00
cron Fix unsafe gateway media path delivery 2026-05-23 01:40:35 -07:00
e2e refactor(gateway): migrate Discord adapter to bundled plugin (full Teams parity) 2026-05-22 14:21:41 -07:00
fakes
gateway fix(dingtalk): finalize open streaming cards before disconnect 2026-05-23 20:48:56 -07:00
hermes_cli feat(kanban): --ids bulk promote + AUTHOR_MAP entry for #29464 2026-05-23 23:10:36 -07:00
hermes_state feat(session_search): single-shape tool with discovery, scroll, browse — no LLM (#27590) 2026-05-17 23:28:45 -07:00
honcho_plugin chore: ruff auto-fix PLR6201 resweep — tuple → set in membership tests (#27355) 2026-05-17 02:29:41 -07:00
integration refactor(gateway): migrate Discord adapter to bundled plugin (full Teams parity) 2026-05-22 14:21:41 -07:00
openviking_plugin
plugins fix(opencode-go): emit Kimi reasoning_effort, match KimiProfile shape 2026-05-23 02:20:28 -07:00
providers fix(custom): pass custom provider extra body 2026-05-21 07:48:53 -07:00
run_agent fix(compressor): ABC compliance — total_tokens, api_mode, logger consistency 2026-05-23 17:38:19 -07:00
scripts feat(acp-registry): switch to uvx distribution, drop npm launcher 2026-05-14 22:27:09 -07:00
skills fix(skills): add timeout to Google OAuth urlopen calls 2026-05-19 00:11:44 -07:00
stress docs: align kanban readiness docs and smoke tests 2026-05-18 21:07:03 -07:00
tools fix(profiles): cross-profile soft guard on file-write tools + system-prompt hint (#31290) 2026-05-24 00:38:17 -07:00
tui_gateway chore: ruff auto-fix PLR6201 resweep — tuple → set in membership tests (#27355) 2026-05-17 02:29:41 -07:00
website
__init__.py
conftest.py test: keep tirith checks hermetic 2026-05-23 02:20:14 -07:00
run_interrupt_test.py
test_account_usage.py
test_atomic_replace_symlinks.py
test_base_url_hostname.py
test_batch_runner_checkpoint.py
test_bitwarden_secrets.py feat(secrets): Bitwarden Secrets Manager integration with lazy bws install (#30035) 2026-05-21 14:10:34 -07:00
test_cli_file_drop.py
test_cli_manual_compress.py fix(tests): catch up six stale tests after compression/aux/kanban changes (#28465) 2026-05-18 21:43:59 -07:00
test_cli_skin_integration.py
test_ctx_halving_fix.py fix(cache): kill long-lived prefix layout — system prompt is now byte-static within a session (#24778) 2026-05-12 20:46:04 -07:00
test_empty_model_fallback.py
test_env_loader_secret_sources.py feat(secrets): label detected credentials with their source (Bitwarden) (#30364) 2026-05-22 03:32:58 -07:00
test_evidence_store.py
test_gateway_streaming_nested_config.py fix(gateway): load streaming config from nested gateway.streaming key 2026-05-14 14:51:07 -07:00
test_get_tool_definitions_cache_isolation.py
test_hermes_bootstrap.py fix(entry-points): guard hermes_bootstrap import so partial updates don't brick hermes (#22091) 2026-05-08 14:43:13 -07:00
test_hermes_constants.py fix(security): guard os.chmod(parent) against / and top-level dirs 2026-05-20 22:56:55 -07:00
test_hermes_home_profile_warning.py
test_hermes_logging.py fix(tests): catch up 25 stale tests after recent merges (#28626) 2026-05-19 01:28:32 -07:00
test_hermes_state.py fix(gateway): separate observed Telegram group context 2026-05-23 01:33:42 -07:00
test_hermes_state_wal_fallback.py fix(sqlite): fall back to journal_mode=DELETE on NFS/SMB/FUSE (#22043) 2026-05-09 02:09:35 -07:00
test_honcho_client_config.py
test_install_sh_browser_install.py fix(install): support non-sudo service-user installs on apt distros (#25814) 2026-05-14 09:05:31 -07:00
test_install_sh_pythonpath_sanitization.py
test_install_sh_setup_wizard_tty_probe.py
test_install_sh_symlink_stomp.py fix(install): preserve pip entry point when re-running on symlinked install 2026-05-14 07:08:45 -07:00
test_install_sh_termux_network_prereqs.py fix: strengthen termux install network prerequisites 2026-05-07 13:04:08 -07:00
test_ipv4_preference.py
test_lazy_session_regressions.py
test_lint_config.py lint: enable PLW1514 as a blocking ruff rule 2026-05-08 14:27:40 -07:00
test_live_system_guard_self_test.py chore: ruff auto-fix PLR6201 resweep — tuple → set in membership tests (#27355) 2026-05-17 02:29:41 -07:00
test_mcp_serve.py fix(mcp): unwrap platforms key in channels_list 2026-05-07 13:41:16 -07:00
test_mini_swe_runner.py
test_minimax_model_validation.py
test_minimax_oauth.py fix(minimax-oauth): refresh short-lived access tokens per request (#30619) 2026-05-22 15:16:15 -07:00
test_minisweagent_path.py
test_model_picker_scroll.py
test_model_tools.py chore: remove Atropos RL environments and tinker-atropos integration (#26106) 2026-05-15 10:36:38 +05:30
test_model_tools_async_bridge.py
test_ollama_num_ctx.py
test_package_json_lazy_deps.py fix(update): make Camofox lazy-installed instead of eager (#27055) 2026-05-16 12:15:45 -07:00
test_packaging_metadata.py
test_plugin_skills.py
test_process_loop_event_loop_warning.py fix(cli): replace get_event_loop() with get_running_loop() to silence RuntimeWarning in process_loop thread (#19285) 2026-05-07 06:35:54 -07:00
test_project_metadata.py fix(packaging): ship dashboard plugin assets in wheel 2026-05-18 20:35:00 -07:00
test_retry_utils.py
test_run_tests_parallel.py test: use subprocesses for each test file (#29016) 2026-05-21 16:40:04 +05:30
test_sanitize_tool_error.py security: sanitize tool error strings before injecting into model context (#26823) 2026-05-16 00:57:39 -07:00
test_sql_injection.py
test_subprocess_home_isolation.py fix: avoid process-wide cron profile home mutation 2026-05-18 17:39:50 +00:00
test_termux_all_extra_compat.py fix: add termux-all install profile and safe fallbacks 2026-05-07 13:04:08 -07:00
test_timezone.py chore: ruff auto-fix PLR6201 resweep — tuple → set in membership tests (#27355) 2026-05-17 02:29:41 -07:00
test_toolset_distributions.py
test_toolsets.py test(toolsets): lock web search into default platform coverage 2026-05-14 08:03:33 -07:00
test_trajectory_compressor.py
test_trajectory_compressor_async.py
test_transform_llm_output_hook.py
test_transform_tool_result_hook.py
test_tui_gateway_server.py fix(tui): stop slash dropdown from chopping last char of /goal (#31311) 2026-05-23 22:12:55 -07:00
test_utils_truthy_values.py
test_yuanbao_integration.py
test_yuanbao_markdown.py
test_yuanbao_pipeline.py
test_yuanbao_proto.py