The cherry-pick comment referenced 'line ~6771' for the /stop handler,
but on current main the handler is at a different offset. Remove the
hard-coded line number — the 'above' reference is sufficient.
17 new tests in tests/gateway/test_subagent_protection_30170.py pin
down both the detection helper and the demotion behaviour:
* TestAgentHasActiveSubagents — 11 cases covering the precision and
defensiveness of _agent_has_active_subagents:
- returns False for None, _AGENT_PENDING_SENTINEL, and stub
agents that lack the _active_children attribute;
- returns False for an empty list (the steady state of an idle
AIAgent);
- returns True for one or many children;
- works when _active_children_lock is None (test stubs);
- rejects truthy MagicMock auto-attributes — this is the
regression-guard for "every MagicMock-based gateway test
suddenly demotes to queue mode" (which is how this was
originally found);
- accepts list/tuple/set as the children container.
* TestBusyHandlerDemotesInterruptForSubagents — 6 cases driving
_handle_active_session_busy_message directly:
- parent.interrupt is NOT called when subagents are active,
message is still merged into the pending queue;
- ack copy mentions "Subagent working", "queued", and the
/stop escape hatch — and does NOT mention "Interrupting";
- with no subagents, behaviour is byte-identical to the
pre-#30170 interrupt path (parent.interrupt called with the
user text, ack says "Interrupting");
- configured queue mode keeps its vanilla "Queued for the next
turn" ack (the #30170 demotion-specific copy must NOT fire);
- configured steer mode still routes to running_agent.steer()
even when subagents are active (the guard is interrupt-only);
- _AGENT_PENDING_SENTINEL does not trigger demotion.
Refs #30170.
When a user sends a conversational follow-up while delegate_task is
running, gateway/run.py calls running_agent.interrupt(event.text) on
the PARENT agent. AIAgent.interrupt() then cascades synchronously
through self._active_children and calls interrupt() on every child
subagent, aborting in-flight delegate_task work. The user sees the
fallback cascade with no root-cause in the gateway log, and minutes of
subagent progress are destroyed — the exact failure mode reported in
Add GatewayRunner._agent_has_active_subagents(running_agent) — a
static helper that returns True iff the parent is currently driving
subagents via delegate_task. The helper is type-defensive: it ignores
truthy MagicMock auto-attributes (so this doesn't accidentally fire
in every test mock that hits the busy path), the _AGENT_PENDING_SENTINEL
placeholder, and missing locks.
Wire the helper into both interrupt branches:
1. _handle_active_session_busy_message — the adapter-level busy
handler. When busy_input_mode == 'interrupt' AND the parent has
active subagents, demote to 'queue' semantics: skip the
parent.interrupt() call, merge the message into the pending
queue, and surface a dedicated ack ("⏳ Subagent working — your
message is queued for when it finishes (use /stop to cancel
everything).") so the operator knows the message wasn't lost and
discovers the explicit escape hatch.
2. The PRIORITY interrupt branch inside _handle_message — the
non-command fast path. Same rationale, same demotion. Routes
through _queue_or_replace_pending_event so the next-turn pickup
stays unchanged.
Explicit /stop and /new commands take a completely different path
(_interrupt_and_clear_session in the slash-command dispatch at line
~6771) and are NOT affected by this guard — the operator still has a
way to force-cancel everything when they actually mean it. Configured
'queue' and 'steer' modes are also untouched: 'queue' already does the
right thing, and 'steer' goes through running_agent.steer() which does
NOT cascade to children (so subagents survive a steer too).
This is Phase 1 of the fix outlined in #30170 — the minimum viable
change that stops subagent loss. Phase 2 (delegation-aware steer
forwarding to active children) and Phase 3 (async delegation, #11508)
are intentionally out of scope.
Refs #30170.
* fix(tui): delineate assistant responses from details
Add a muted Response marker before assistant text when thinking/tool details are visible so reasoning and final output do not visually run together.
* fix(tui): account for response separator height
Keep virtual transcript estimates aligned with the new response separator and avoid allocating trimmed copies of long assistant text.
* fix(tui): gate response separator estimate on details
Only add response-separator height when assistant details actually render, and use a non-allocating body-text check.
* fix(tui): skip empty detail height estimates
Do not add virtual transcript height for assistant details when no thinking or tool detail UI will render.
* fix(tui): estimate details by section visibility
Pass resolved thinking/tool visibility into virtual height estimates so hidden detail sections do not reserve response-separator rows.
After key #1 is marked exhausted the retry still called the API with key #1
due to env-var bias in _get_cached_client / resolve_api_key_provider_credentials.
Fix: peek the pool and pass the active entry's key as explicit_api_key.
Secondary: api_key_hint in mark_exhausted_and_rotate pins the correct entry
under concurrent CLI+gateway calls; _is_payment_error matches GoUsageLimitError;
extract_api_error_context parses "Resets in Xhr Ymin".
Adds two new config keys:
- paste_collapse_threshold (default: 5) — line count threshold for
bracketed paste collapse in both TUI and CLI
- paste_collapse_threshold_fallback (default: 0, disabled) — same for
the fallback heuristic in terminals without bracketed paste support
TUI frontend reads these from config.get full via applyDisplay/patchUiState.
CLI reads from self.config at paste-handling time.
Closes#5626
Related: #5623
Closes#26145.
When the user interrupts the retry loop between two 429s (Ctrl-C in
interactive mode, /new, gateway disconnect), the local has_retried_429
flag dies with the recovery function. On the next user prompt the agent
restarts with has_retried_429=False, hits 429 on the exhausted credential,
sets the flag, returns 'retry once'. Repeat forever — the second 429 that
would trigger rotation is never reached, and healthy entries (priority>0
free/paid accounts) are never tried.
Fix: in recover_with_credential_pool's rate_limit branch, pre-check
pool.current().last_status before running the retry-once dance. If the
current entry is already STATUS_EXHAUSTED, rotate immediately. Uses
getattr() for the attribute read so existing tests with SimpleNamespace
mocks (which only set 'label') keep working.
Co-authored-by: zccyman <16263913+zccyman@users.noreply.github.com>
The new install-path validator from this PR raises 'Unsafe install path:
...' earlier in the pipeline than the previous resolve-then-check path.
Behavior is identical (ok=False, victim untouched, refused before
rmtree) — only the error string changed.
Validate Skills Hub lock-file install paths at both ends of the
lifecycle so a poisoned or malformed lock.json entry cannot drive
shutil.rmtree to a location outside SKILLS_DIR:
- HubLockFile.record_install rejects empty/'.'/absolute/traversal/
Windows-drive paths at write time, and requires the final path
component to match the skill name (shape: '<skill>' or
'<category>/<skill>').
- install_from_quarantine resolves its destination through the same
validator, catching symlink/junction redirects inside skills/.
- uninstall_skill resolves the lock entry through the new validator
before rmtree. Refuses anything that resolves to SKILLS_DIR itself
(empty/dot paths) or to a target outside SKILLS_DIR (absolute paths,
traversal, symlinked dirs in skills/ pointing outward).
- 14 focused regression tests covering each rejection class plus a
symlink-redirect case.
E2E verified: hand-crafted poisoned lock.json entries (absolute path,
empty install_path, traversal) all refuse and leave the targeted
victim untouched; legitimate uninstall still succeeds.
Co-authored-by: Teknium <127238744+teknium1@users.noreply.github.com>
Nous Portal is OAuth-only (auth_type=oauth_device_code, no API key path),
but the non-retryable-401 guidance branch only covered openai-codex and
xai-oauth. A Nous 401 fell through to the generic 'Your API key was
rejected... run hermes setup' message, which is wrong advice — the user
needs hermes auth add nous --type oauth, not an API key.
Also flag the case where the failing model slug ends in :free (OpenRouter
syntax) while provider is nous. Without that hint, users re-OAuth
successfully and then hit the same 401 on the next message because Nous
Portal doesn't carry the OpenRouter free-tier slug.
Reported by ashh — debug dump showed Nous device_code exhausted +
deepseek/deepseek-v4-flash:free as the model.
Aux callers (title generation, vision, session search, etc.) can reach
resolve_provider_client() without an explicit model when the user
picked their main provider via 'hermes model' and didn't bother
configuring a per-task auxiliary.<task>.model override. The
expectation in that case is universal: 'use my main model for side
tasks too.'
Before, the OAuth providers (xai-oauth, openai-codex) silently
returned (None, None) on an empty model — both lack a catalog default
because their accepted-model lists drift on the backend. That caused
_resolve_auto to drop to its Step-2 fallback chain (OpenRouter /
Nous / etc.), so aux tasks billed against the wrong subscription
without warning.
The fix is at the top of resolve_provider_client() — a single
3-step universal fallback that runs before any provider branch, so
no provider-specific empty-model guards are needed (now or for any
future provider we add):
1. caller-passed model (caller knew what they wanted)
2. provider's catalog default (cheap aux model, if registered)
3. user's main model from config.yaml
Behaviour by provider class:
- OAuth providers (xai-oauth, openai-codex) — no catalog default, so
step 3 applies. Title gen runs on grok-4.3 / gpt-5.4 against the
user's actual subscription instead of leaking to OpenRouter.
- API-key providers (anthropic, gemini, kimi-coding, etc.) — catalog
default wins at step 2, preserving the original 'cheap aux model'
behaviour. Anthropic users still get claude-haiku-4-5 for titles,
not opus.
- Explicit-model callers (auxiliary.<task>.model config, programmatic
callers) — caller wins at step 1, no surprise switching.
Salvaged from @wysie's PR #31845 which fixed the xai-oauth branch
specifically. The universal shape supersedes the per-branch fix
and covers openai-codex (same bug class) plus any future OAuth
providers.
4 new tests in TestResolveProviderClientUniversalModelFallback:
- empty_model_for_oauth_provider_falls_back_to_main_model
- empty_model_for_codex_also_uses_main_model
- empty_model_for_catalog_provider_uses_catalog_default
- explicit_model_takes_precedence_over_fallbacks
365/365 across tests/agent/test_auxiliary_*, tests/run_agent/test_codex_xai_oauth_recovery.py, tests/hermes_cli/test_auth_xai_oauth_provider.py, and tests/hermes_cli/test_plugin_auxiliary_tasks.py.
Co-authored-by: wysie <wysie@users.noreply.github.com>
All four failures were broken by the security cluster (#10082 / #10133 /
#4609 / symlink-reject batch) merging on May 25. They were red on
origin/main HEAD when #32042 and #32061 ran, gating PRs that touched
unrelated code.
1) tests/hermes_cli/test_update_zip_symlink_reject.py
test_update_via_zip_accepts_normal_member called the real
_update_via_zip without sandboxing PROJECT_ROOT — so the function's
shutil.copytree() actually copied the fake README from the test ZIP
over the real repo's README.md, which then made
test_readme_mentions_powershell_installer fail in any test run that
happened to pick this test up earlier. Mock PROJECT_ROOT to an
isolated tmp_path / install_dir, stub subprocess so pip/uv reinstall
doesn't actually run, and assert the fake README lands in the
sandbox (not the real tree).
2) tests/tools/test_windows_native_support.py
test_readme_mentions_powershell_installer was the victim of (1) —
nothing wrong with the test itself, the fix in (1) clears it.
3) tests/tools/test_file_read_guards.py
test_proc_fd_other_not_blocked called _is_blocked_device('/proc/self/fd/3')
expecting False. But _is_blocked_device runs realpath() and on
pytest xdist workers fd 3 happens to be dup'd to /dev/urandom
(because the worker subprocess inherits open fds from pytest's
collection pipe machinery). Switch to the lower-level
_is_blocked_device_path which is the path-pattern check the test
actually means to exercise; realpath-resolution coverage already
lives in test_symlink_to_blocked_device_is_blocked.
4) tests/tools/test_transcription_tools.py
Module installed a faster_whisper stub via sys.modules without
setting __spec__, then later @pytest.mark.skipif called
importlib.util.find_spec('faster_whisper') which raises
'ValueError: __spec__ is None' for modules with a None spec attr.
Set __spec__ on the stub to a real ModuleSpec.
Validation: 195/195 green across the 4 affected files.
The TUI frontend's slash command registry shadowed /queue's 'q' alias
with /quit's 'q' alias. Since /quit appeared later in the registry,
the flat lookup kept the later entry, making /q always quit instead
of queueing a prompt.
This mirrors the backend fix in PR #10538 (hermes_cli/commands.py)
but applies the same correction to the TUI TypeScript registry.
Fixes#10467
When an MCP server triggers OAuth at startup, the user can now type 'skip'
(or 'cancel', 's', 'n', 'no', 'q', 'quit') at the paste prompt + Enter to
exit the flow cleanly and continue agent startup without that server.
Previously the only ways to bypass an unwanted OAuth prompt were:
- Wait the full 5-minute paste timeout
- Ctrl+C (also kills the whole reload, may leave half-state)
- Edit config.yaml to set 'enabled: false' on the server
Skip writes a sentinel to result['error'] which _wait_for_callback maps to
OAuthNonInteractiveError('user_skipped'). mcp_tool already classifies that
as an auth error in _is_auth_error() and the reconnect loop logs it as
'not retrying automatically' — server stays disconnected for the session,
other MCP servers continue normally, no infinite retry burn.
The skip message tells users how to re-auth later ('hermes mcp login') or
disable persistently ('enabled: false'), so they don't have to remember.
14 new tests covering: case-insensitive skip parsing, all 7 skip tokens,
skip not stomping an HTTP-listener win, skip routed to skip path rather
than URL-parse path, sentinel mapped to OAuthNonInteractiveError, prompt
mentions the skip option.
Follow-up to #32053. The OAuth-over-SSH guide and the MCP feature page
previously only covered xAI and Spotify. Now that MCP servers can complete
OAuth via stdin paste-back on remote/headless hosts, document it.
oauth-over-ssh.md:
- Add MCP servers to the 'Which Providers Need This' table.
- New 'MCP Servers' section covering: paste-back (no setup, works
anywhere), SSH port forward (same pattern as xAI/Spotify), and the 30s
config-auto-reload race pitfall (use 'hermes mcp login <server>' from a
fresh terminal instead of editing config from inside a running session).
mcp.md:
- New 'OAuth-authenticated HTTP servers' section under HTTP servers,
covering auth: oauth config, token cache path, paste-back vs SSH
tunnel for headless hosts, and the same reload-race pitfall.
- Cross-links to the OAuth-over-SSH guide anchor.
The CLI status bar tracked /background agent tasks (▶ N) but not shell
processes spawned via terminal(background=true). Both kinds of work can
run concurrently and a user has no in-bar signal for shell processes.
Add an independent indicator (⚙ N) sourced from
tools.process_registry.process_registry._running. The two indicators
render side-by-side when both are active (▶ 1 │ ⚙ 2), hidden when their
count is zero. Renders at all four status-bar tiers (text fallback +
prompt_toolkit fragments, narrow + wide widths). The narrow <52 tier
still drops both for space — unchanged.
New ProcessRegistry.count_running() returns len(_running) without
acquiring _lock; CPython dict len is atomic and we're polling on every
status-bar tick, so lock-free is the right tradeoff.
The chatgpt.com/backend-api/codex endpoint has an intermittent failure mode
where it accepts the connection but never emits a single stream event — the
socket just hangs. Direct sequential probing reproduces it (0 events, no HTTP
status), and a fresh reconnect then succeeds in ~2s. Today the only guard is
the wall-clock stale timeout in interruptible_api_call, so a dead-on-arrival
connection is held for the full stale window (90-900s depending on context /
config) before the retry loop can reconnect — minutes of wasted wall time per
stall, at a rate of ~20% of calls during affected windows.
Add a TTFB watchdog scoped to the codex_responses path:
- codex_runtime.run_codex_stream stamps agent._codex_stream_last_event_ts on
*every* stream event (not just output-text deltas), so reasoning-only and
tool-call-only turns are not mistaken for a stall.
- interruptible_api_call resets that marker before the worker starts and, while
it is still None, kills the connection once elapsed exceeds the TTFB cutoff
(default 45s, tunable via HERMES_CODEX_TTFB_TIMEOUT_SECONDS, 0 disables). The
raised TimeoutError flows through the existing retry path unchanged.
Once any event has arrived the stream is healthy and only the existing
wall-clock stale timeout applies, so legitimate long generations are never
interrupted. Gated to codex_responses; the chat_completions non-stream,
anthropic and bedrock branches have no first-event signal and are untouched.
Adds tests/agent/test_codex_ttfb_watchdog.py covering the stall kill, the
events-flowing pass-through, and the env-disable path.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
The gateway's media delivery allowlist required files live inside
`~/.hermes/cache/{documents,images,...}`, which is the wrong shape for
real agent usage. Agents naturally produce artifacts via terminal tools
(`pandoc -o /tmp/report.pdf`, `matplotlib savefig`, etc.) or
write_file into project directories — these never land under the cache.
Result: users got a raw file path in chat instead of an attachment.
This is doubly bad in deployment shapes where the cache directories
aren't writable by the agent at all: Hermes running in Docker with a
read-only mount, or with a Docker/Modal/SSH terminal backend whose
filesystem isn't the gateway host's filesystem.
Layered trust model:
1. Cache-dir allowlist (unchanged) — Hermes-managed roots always trusted.
2. Operator allowlist — `HERMES_MEDIA_ALLOW_DIRS` env var, now also
surfaced as `gateway.media_delivery_allow_dirs` in config.yaml.
3. Recency-based trust (new, default on) — files whose mtime is within
`gateway.trust_recent_files_seconds` (default 600s) of "now" are
trusted even outside the cache/operator allowlist. Old host files
(`/etc/passwd`, `~/.bashrc`, `~/.ssh/id_rsa`) have mtimes measured
in days/months, well outside the window — prompt-injection paths
pointing at pre-existing files are still rejected.
4. Hard denylist — `/etc`, `/proc`, `/sys`, `/dev`, `/root`, `/boot`,
`/var/{log,lib,run}`, plus `$HOME/.{ssh,aws,gnupg,kube,docker,config,
azure,gcloud}` and `Library/Keychains`. Denylist blocks delivery
even when recency would trust the file, in case an attacker
somehow refreshes a sensitive file's mtime.
Operators who want strict-allowlist behavior set
`gateway.trust_recent_files: false` and the system reverts to
pre-existing behavior.
Tests: 6 new cases in test_platform_base.py cover the recency window,
disabled mode, system-path denylist, and the motivating PDF-in-project
scenario. 3 existing tests (test_platform_base, test_tts_media_routing,
test_send_message_tool) that exercised the strict-allowlist path are
updated to disable recency trust explicitly.
E2E validation: real `validate_media_delivery_path()` accepts fresh
PDFs in /tmp and project dirs, rejects /etc/passwd, ~/.ssh/id_rsa, and
files older than the window; config.yaml `gateway.*` keys bridge
correctly to the env vars the validator reads.
When the user runs OAuth on a remote/SSH machine without a port forward,
the OAuth provider redirects to http://127.0.0.1:<port>/callback which
only the listener on the remote machine can receive — the user's browser
on another box just shows a connection error.
_wait_for_callback() now races the HTTP listener against a stdin reader
on interactive TTYs. The user can copy the URL from the browser's address
bar after authorization (which contains code=...&state=...) and paste it
back at the prompt. Whichever fills the result dict first wins; the HTTP
listener remains the primary path for local sessions and SSH tunnels.
Accepts any of:
- Full local redirect URL: http://127.0.0.1:N/callback?code=...&state=...
- Provider URL after redirect: https://mcp.linear.app/callback?code=...&state=...
- Just the query string: ?code=...&state=... or code=...&state=...
The paste thread only spawns when _is_interactive() is true, preserving
the existing 'no input() in headless runs' invariant — verified by
TestWaitForCallbackPasteIntegration.test_paste_prompt_NOT_shown_when_noninteractive.
The SSH-session hint in _redirect_handler is updated to surface the paste
option as the primary remedy, with ssh -L tunneling as the alternative.
_update_via_zip downloads a source ZIP from GitHub and calls
zipfile.ZipFile.extractall. The existing zip-slip path guard validates
each member's path stays under tmp_dir, but does not check member type
— so a ZIP containing a symlink member would still be materialized by
extractall, and a symlink target could point outside the extracted
tree (or to a sensitive system path).
This isn't a high-likelihood threat for hermes-agent's actual GitHub
source ZIPs (we don't ship symlinks), but the extractall path runs as
the user's account and a compromised mirror could plant arbitrary files
via the symlink → target → write chain.
Reject any member whose Unix mode bits (upper 16 bits of external_attr)
are S_IFLNK before extractall. Hermes source ZIPs contain only regular
files and directories; a symlink member is unambiguously suspicious.
Regression tests cover: symlink member rejection (raises ValueError,
caught by the outer try/except as a clean SystemExit, no extraction),
and the happy-path verification that a normal ZIP doesn't trigger the
symlink reject message.
Salvaged from PR #15881 by @codeblackhole1024. The remaining pieces of
that PR were already on main or contradicted explicit design decisions:
- config.yaml write-deny: already in agent/file_safety.py's
control_file_names denylist (the modern guard); the proposed addition
to build_write_denied_paths was the legacy path.
- Quick commands danger detection: contradicts the explicit
cli.py:8491-8492 comment 'shell=True is intentional: quick_commands
are user-defined shell snippets from config.yaml — not agent/LLM
controlled.'
- Memory plugin shlex.split for dep checks: already on main
(hermes_cli/memory_setup.py:133).
Co-authored-by: teknium1 <127238744+teknium1@users.noreply.github.com>
text_to_speech_tool accepts an explicit output_path. Without a traversal
guard, a path containing '..' components (whether prompt-injection-
controlled, from a confused skill, or just a buggy caller) could escape
its declared base and write the audio to a system location — e.g.
`output_path='audio/../../etc/cron.d/x'` lands the file outside the
intended audio cache.
Reject '..' components in the user-supplied path. Explicit absolute
paths are unchanged (the agent legitimately writes audio wherever the
user/caller asks); only traversal-style escapes are blocked. The
terminal tool can still write anywhere with approval — this just keeps
the unattended TTS surface from materializing files via traversal.
Regression tests cover: '..' in the middle (audio/../../etc/...),
bare '..' prefix, and the negative cases (absolute paths + relative
paths without '..' both pass through unchanged).
Salvaged from PR #6693 by @aaronlab. The original PR confined output to
DEFAULT_OUTPUT_DIR-or-cwd, which broke 9 existing tests that legitimately
write to tmp_path locations. The traversal-only check covers the actual
threat (path-escape via '..' from prompt injection) without restricting
where users can choose to write their audio.
The remaining pieces of #6693 (skill_commands rglob symlink rejection,
delegate_tool batch prefix display) are dropped:
- skill_commands rglob: breaks the documented design supporting
~/.hermes/skills/<name> as a symlink to a checked-out skill elsewhere
(see comment at agent/skill_commands.py:73-75)
- delegate_tool batch prefix: pure UX, doesn't belong in a security PR
Co-authored-by: teknium1 <127238744+teknium1@users.noreply.github.com>
* fix(streaming): route mid-tool-call partial-stream-stub through length continuation (#31998)
When a stream stalls mid-tool-call (e.g. a large write_file), the
partial-stream-stub recovery used finish_reason='stop' which caused the
conversation loop to treat the turn as complete, returning only the
warning text. When users said 'continue', the model retried the same
large tool call, hit the same stale timeout, and looped indefinitely.
Changes:
- chat_completion_helpers.py: change _stub_finish_reason from 'stop' to
'length' for mid-tool-call partials. The stub still has tool_calls=None
so no tool auto-executes — the model gets a fresh API call through the
existing length-continuation machinery (bounded to 3 retries).
Also attach _dropped_tool_names to the stub for downstream use.
- conversation_loop.py: add a third continuation prompt branch for
partial-stream-stubs with dropped tool calls. Instead of the generic
'continue where you left off' (which would retry the same large call),
tell the model to break the output into smaller tool calls (~8K
tokens each) to avoid stream timeouts.
- test_partial_stream_finish_reason.py: update existing test from
finish_reason='stop' to 'length', add _dropped_tool_names assertion,
add new test_dropped_tool_call_uses_chunking_prompt for the 3-way
prompt branching.
Safety: tool_calls=None is preserved on the stub, so the conversation
loop enters the text-continuation branch (line 1513), NOT the tool-call
execution branch (line 3246). No tool auto-executes. The model simply
gets another API call with targeted guidance.
* refactor: extract constants and continuation prompt helper
- Move magic strings to hermes_constants.py (PARTIAL_STREAM_STUB_ID,
FINISH_REASON_LENGTH)
- Extract _get_continuation_prompt() in conversation_loop.py — DRYs the
3-way prompt branching and lets tests import the real function
- Trim verbose inline comments in chat_completion_helpers.py
- Tests import constants + helper instead of duplicating logic
---------
Co-authored-by: alt-glitch <balyan.sid@gmail.com>
The test set HERMES_YOLO_MODE=1 via monkeypatch.setenv, expecting
check_dangerous_command() to honor yolo and bypass cron_mode=deny. But
tools.approval._YOLO_MODE_FROZEN is intentionally frozen at module
import time (security: prevents prompt-injection runtime escalation).
When CI imports the module BEFORE the test sets the env, the frozen
value stays False and the yolo bypass never activates.
Local runs missed this because the conftest leaked a non-empty
HERMES_YOLO_MODE into the import-time env. CI's clean-env path exposed
the bug deterministically on test (3) / test (4) shards.
Fix: patch the module attribute directly via mock.patch.object so the
test simulates process-startup-with-yolo regardless of import order.
The behavior under test (yolo bypasses cron_mode=deny for non-hardline
commands) is unchanged; the security invariant (_YOLO_MODE_FROZEN can't
be set at runtime by skills) is preserved.
Reproduced locally with: env -i HOME=$HOME PATH=$PATH python3 -m pytest
tests/tools/test_cron_approval_mode.py -o 'addopts=' -v
Without the fix: 1 failed, 23 passed. With the fix: 24 passed.
* fix(transcription): reject symlinked audio inputs
Validation runs before provider selection, so rejecting symbolic-link paths there prevents supported-extension links from being treated as normal audio files. Use os.path.islink to avoid perturbing the existing Path.stat error path and to reject links before resolving targets.
Constraint: Keep validation platform-safe and avoid requiring symlink support where unavailable.
Rejected: Use Path.is_symlink | it consumes pathlib stat calls and broke the existing stat error regression.
Confidence: high
Scope-risk: narrow
Directive: Keep path hardening in _validate_audio_file before provider dispatch.
Tested: source venv/bin/activate && python -m pytest tests/tools/test_transcription_tools.py::TestValidateAudioFileEdgeCases -q (5 passed)
Tested: source venv/bin/activate && python -m pytest tests/tools/test_transcription_tools.py::TestValidateAudioFileEdgeCases tests/tools/test_transcription_tools.py::TestTranscribeAudioDispatch::test_invalid_file_short_circuits -q (6 passed)
Tested: source venv/bin/activate && python -m compileall tools/transcription_tools.py tests/tools/test_transcription_tools.py
Tested: git diff --check
Not-tested: Full tests/tools/test_transcription_tools.py under .[dev] only; existing faster_whisper optional dependency tests fail with ModuleNotFoundError.
* Keep transcription tests independent of optional whisper install
The transcription suite mocks faster-whisper directly, so a minimal test stub keeps the branch verifiable in environments where the optional package is not installed. This preserves the existing mock-based coverage without adding a dependency.
Constraint: faster-whisper is an optional local STT dependency and is absent from the current validation environment
Rejected: Install faster-whisper just for branch validation | would add heavyweight environment coupling outside the patch scope
Confidence: high
Scope-risk: narrow
Directive: Keep this as a test-only stub unless production import semantics change
Tested: pytest tests/tools/test_transcription_tools.py -q
---------
Co-authored-by: WuKongAI-CMU <210765158+WuKongAI-CMU@users.noreply.github.com>
* fix: reject read_file symlinks to blocking devices
The read_file guard already refused direct device paths such as /dev/zero, but a workspace symlink resolving to one of those devices could still reach the shell-backed read path and hang on wc/head/sed. Keep the literal alias check and add a resolved-path pass so local symlinks to blocked device/fd endpoints are rejected before I/O.
Constraint: Preserve literal /dev/stdin handling before terminal-specific realpath resolution
Confidence: high
Scope-risk: narrow
Tested: pytest tests/tools/test_file_read_guards.py tests/tools/test_file_tools.py -q; python -m compileall tools/file_tools.py tests/tools/test_file_read_guards.py; git diff --check
Signed-off-by: WuKongAI-CMU <210765158+WuKongAI-CMU@users.noreply.github.com>
* Keep file guard tests off sensitive macOS temp paths
The branch now inherits a sensitive-path write guard from upstream main. On macOS, tempfile.mkdtemp() resolves under /private/var/folders, so the new write-path guard fired before the file read dedup assertions could exercise their intended behavior. The tests now create their scratch files inside the worktree temp checkout, outside those system-sensitive prefixes, without changing production behavior.
Constraint: Rebased branch must pass the expanded file read guard suite on macOS.
Rejected: Loosen the production sensitive-path prefix list | broader behavior change unrelated to this PR.
Confidence: high
Scope-risk: narrow
Tested: pytest tests/tools/test_file_read_guards.py -q
---------
Signed-off-by: WuKongAI-CMU <210765158+WuKongAI-CMU@users.noreply.github.com>
Co-authored-by: WuKongAI-CMU <210765158+WuKongAI-CMU@users.noreply.github.com>
The read_file tool and terminal cat can access /proc/self/environ to
recover all process env vars including secrets stripped by the subprocess
blocklist. Output redaction partially mitigates (catches known-format
tokens) but misses custom/proprietary key formats, especially when
values are printed without their key names.
Add /proc/*/environ, /proc/*/cmdline, and /proc/*/maps to the blocked
device paths in _is_blocked_device():
- /proc/*/environ: leaks full process env (API keys, tokens)
- /proc/*/cmdline: leaks command-line args (may contain passwords)
- /proc/*/maps: leaks memory layout (ASLR bypass for exploitation)
Legitimate /proc reads (cpuinfo, meminfo, uptime, version) remain
accessible — the check only blocks per-pid pseudo-files with known
sensitive suffixes.
Complements PR #4432 (PID namespace isolation for child processes)
which prevents children from reading the parent's /proc, but does not
prevent the parent process itself from being read via file tools.
Partially addresses #4427
Changes:
tools/file_tools.py | +6
tests/tools/test_file_read_guards.py | +18 -1
Co-authored-by: dsr-restyn <dsr-restyn@users.noreply.github.com>
When the terminal drops the ESC[201~ end mark during a bracketed paste
(terminal race, torn write, SSH glitch, macOS sleep/wake), prompt_toolkit's
Vt100Parser keeps buffering all later input in _paste_buffer forever. From
the user's perspective, the CLI appears frozen — the only recovery was
closing the tab/session.
This patch monkey-patches Vt100Parser.feed() so that bracketed-paste mode
flushes buffered content as a normal BracketedPaste event after 2 seconds
without an end marker, then restores normal parsing.
Includes 8 regression tests covering normal paste, timeout recovery,
torn end marks, and edge cases.
Surgical reapply of PR #27518. Original branch was many months stale
(1193 files / 172k LOC of unrelated reverts); the substantive ~77 LOC
patch in cli.py plus the new 157-line test file were reapplied onto
current main with the contributor's authorship preserved via --author.
A Ctrl+C during a slow slash command (e.g. /skills browse on a large
skill tree, /sessions list against a multi-GB SQLite DB) used to unwind
past self.process_command() to the outer prompt_toolkit event loop,
which killed the entire session — losing all conversation state.
Fix: wrap the slash-command dispatch in try/except KeyboardInterrupt
so Ctrl+C aborts the command but the prompt loop continues. Other
exceptions still propagate so real bugs aren't silently swallowed.
Surgical reapply of PR #5189. Original branch was many months stale
(3764 files / 1M+ LOC of unrelated reverts); the substantive ~6 LOC
change in cli.py was reapplied by hand onto current main with the
contributor's authorship preserved via --author.
On Windows (PowerShell/Windows Terminal), the queue-based modal used for
destructive slash command confirmations deadlocks because prompt_toolkit's
input channel becomes unresponsive when entered from the process_loop daemon
thread. Keystrokes never reach the key bindings, so response_queue.get()
blocks until the 120-second timeout expires.
Fix: fall back to _prompt_text_input (stdin-based) when:
1. sys.platform == 'win32' — Windows console doesn't support the modal reliably
2. Called from non-main thread — key bindings can't fire from daemon threads
3. self._app is not set — existing behavior for tests/non-interactive
This mirrors the thread-aware guard from _prompt_text_input (PR #23454).
9 new regression tests covering Windows detection, non-main thread fallback,
macOS/Linux modal preservation, and integration with _confirm_destructive_slash.
Fixes#30768
Surgical reapply of PR #30773. Original branch was many months stale (911
files / 146k LOC of unrelated reverts); the substantive ~30 LOC change in
cli.py plus the new test file were reapplied onto current main with the
contributor's authorship preserved via --author.
The ChatGPT Codex backend (chatgpt.com/backend-api/codex) has historically
silently dropped certain model requests: the connection is accepted but no
stream events are emitted and no error is raised. PR #31967 lowered the
implicit stale-call default from 300s to 90s so fallbacks kick in faster,
but users still see an opaque "No response from provider for 90s
(non-streaming, ...)" message that gives no path forward.
This patch adds a narrow heuristic — gpt-5.5 family on the Codex backend
via codex_responses api_mode — that substitutes the generic timeout
message with actionable text naming the gpt-5.4-codex workaround and
pointing at #21444 for symptom history.
Changes:
- run_agent.py — new ``AIAgent._codex_silent_hang_hint(model=...)`` method.
Returns ``None`` for any request that does not match all three guards
(codex_responses api_mode, openai-codex provider or chatgpt.com Codex
base URL, gpt-5.5-family model name with word-boundary regex anchoring
to avoid false-positives on e.g. ``gpt-5.50``).
- agent/chat_completion_helpers.py — the non-stream stale-call site
consults the hint via ``getattr(...)`` so the call site stays robust
if the helper is ever removed or stubbed in tests. Hint is appended to
both the ``_emit_status`` warning and the ``TimeoutError`` message so
the user sees it in their terminal AND it lands in any retry-loop
diagnostics.
- tests/run_agent/test_codex_silent_hang_hint.py — 10 regression tests
covering positive cases (bare gpt-5.5, vendor-prefixed openai/gpt-5.5,
gpt-5.5-codex SKU, model=None fallback to self.model) and negative
cases (gpt-5.4-codex workaround, gpt-5.50 false-positive guard,
non-codex api_mode, non-codex provider, empty/None model, unrelated
models on Codex).
Does NOT fix the backend-side issue (that's an upstream OpenAI/ChatGPT
problem we cannot patch from here). Only converts an opaque timeout into
text that names the workaround so users do not have to dig through logs
or wait for a forum post to learn what to do.
Closes#22046
get_read_block_error() only blocked internal Hermes cache files but
allowed reading project-local secret-bearing environment files (.env,
.env.production, .env.local, etc.) through both read_file and ACP
fs/read_text_file paths.
Add a basename deny set for common secret-bearing .env variants.
.env.example remains readable as documentation.
Fixes#20734
.env holds API keys and secrets. Multiple creation sites used `cp` /
`touch` / `shutil.copy2` which obey the process umask — commonly
0o022, leaving the file at 0o644 (world-readable). Apply chmod 0o600
explicitly at every site that creates or copies .env.
Sites covered:
- docker/stage2-hook.sh: after the seed_one '.env' call, applied
unconditionally (not just on first-seed) so a host-mounted .env with
loose perms gets tightened on every container restart
- hermes_cli/doctor.py: 'hermes doctor --fix' touches an empty .env
when missing
- hermes_cli/profiles.py: 'hermes profile create --clone' copies .env
from the source profile; shutil.copy2 preserves source mode, so a
source .env at 0o644 was being cloned into 0o644
- setup-hermes.sh: in-tree setup script's cp .env.example .env path,
plus the already-exists branch (mirror of install.sh which already
chmods 600 unconditionally on line 1442)
scripts/install.sh was NOT changed — it already chmod 600's the .env
unconditionally after the create/already-exists branches (line 1442).
Salvaged from PR #25726 by @dusterbloom. The docker/entrypoint.sh
portion of the original PR was dropped because main switched to an
s6-overlay shim — the .env creation logic moved to stage2-hook.sh,
which is where the chmod now lives.
Closes#25497 (subset — install.sh + setup-hermes.sh) and #8448
(subset — install.sh only) as superseded.
Co-authored-by: teknium1 <127238744+teknium1@users.noreply.github.com>
* fix(approval): harden YOLO bypass, LLM parsing, auto-approve audit, pipe pattern
- BUG-009 (CRITICAL): freeze HERMES_YOLO_MODE at module import via
_YOLO_MODE_FROZEN; prevents skills/prompt-injection from calling
os.environ["HERMES_YOLO_MODE"]="true" at runtime to bypass all checks
- BUG-002 (HIGH): replace substring "APPROVE" in answer with exact
answer == "APPROVE" in _smart_approve; prompt already requests exactly
one word, substring match was exploitable via verbose LLM responses
- BUG-001 (MEDIUM): add logger.warning for every dangerous command that
auto-approves in non-interactive non-gateway context; makes silent
approvals visible in audit logs without breaking script behavior
- BUG-008 (LOW): expand curl/wget pipe pattern to cover | /bin/bash and
| bash -c variants, not just | sh / | bash
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
* fix(approval): add missing is_truthy_value import + fix yolo test patches
_YOLO_MODE_FROZEN uses is_truthy_value() from utils — import was missing.
Tests that set HERMES_YOLO_MODE via monkeypatch.setenv() no longer work
because the value is frozen at import time; update them to patch the
module-level flag directly via monkeypatch.setattr().
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
---------
Co-authored-by: Claude Sonnet 4.6 <noreply@anthropic.com>
register_env_passthrough() (the skill-declared path) filters out names in
_HERMES_PROVIDER_ENV_BLOCKLIST and logs a warning citing GHSA-rhgp-j443-p4rf.
_load_config_passthrough() (the config.yaml path) did not. Both feed the
same is_env_passthrough() allowlist that local.py and code_execution_tool.py
consult before stripping a variable from the child env.
A skill that wanted to leak ANTHROPIC_API_KEY or OPENAI_API_KEY into
execute_code could no longer self-register the name (the GHSA fix
blocks it), but the same outcome was still reachable by asking the
operator to add the name to terminal.env_passthrough in config.yaml,
or by any in-process actor with write access to ~/.hermes/config.yaml.
Apply the same _is_hermes_provider_credential filter inside
_load_config_passthrough, mirroring the skill-path warning so operators
see the same explanation. Non-Hermes API keys (TENOR_API_KEY,
NOTION_TOKEN, etc.) are unaffected since they are not in the blocklist.
* perf(bitwarden): persist secret-fetch cache across CLI invocations
Every `hermes` invocation paid a ~380ms tax for `bws secret list` to
Bitwarden Secrets Manager because the existing cache was in-process only.
Back-to-back `hermes chat -q`, gateway-spawned agents, and cron-launched
runs all re-fetched.
Adds a disk-persisted L2 cache at `<hermes_home>/cache/bws_cache.json`
(mode 0600, never contains the access token — only the SHA-256
fingerprint prefix). Same TTL as the in-process cache. Read on miss,
write on bws success, ignored on key mismatch / corruption / expiry.
Measured on a startup profile:
load_hermes_dotenv() cold: 372ms → warm (disk cache hit): 20ms
End-to-end `hermes --version` cold→warm: 666ms → ~295ms.
In a hermes-vs-codex benchmark across 11 single- and multi-turn tasks
(framework overhead = wall − llm − tool_exec, median over 3 trials):
cohort before after saved
single-turn (median) 2.96s 2.31s -0.65s
multi-turn (5-turn) 9.40s 8.95s -0.45s (≈0.3s/turn)
Hermes now wins head-to-head on 6/11 tasks vs codex (was 4/11 before).
The remaining ~0.6s single-turn delta is mostly Python's own import
cost in hermes_cli.main, which is a separate optimization.
* perf(cli): lazy-load model catalog + dedupe config.yaml reads at startup
Two import-time wins on top of the bws disk-cache fix:
1. Lazy-load `hermes_cli.models._PROVIDER_MODELS` via PEP 562
module-level `__getattr__`. The catalog is ~55ms of work that was
eagerly imported on every CLI invocation (line 4557 `if not
_is_termux_startup_environment(): from hermes_cli.models import
_PROVIDER_MODELS`). Audit showed every internal call site already
does its own function-local import; only test code reads
`hermes_cli.main._PROVIDER_MODELS` as a module attribute, and
__getattr__ keeps that working transparently. First access triggers
the import once and caches the result on the module via
`globals()[name] = ...`, so subsequent reads are dict lookups.
2. Dedupe the double config.yaml read in the top-of-module bootstrap.
Previously: one raw yaml.safe_load for the `security.redact_secrets`
bridge, then a separate full `load_config()` (with deep-merge) for
`network.force_ipv4`. Both keys come from the same file. Merged
into one raw yaml load.
Combined with the bws cache fix in the previous commit:
hermes --version wall time:
original (cold): 666 ms
after bws fix (warm): 295 ms
after lazy-load + dedupe: 228 ms (-67 ms additional, -66% from original)
Tests:
- tests/hermes_cli/test_api_key_providers.py: 173/173 pass
(lazy __getattr__ correctly handles
`from hermes_cli.main import _PROVIDER_MODELS`)
- tests/test_ipv4_preference.py + tests/hermes_cli/test_redact_config_bridge.py +
tests/agent/test_redact.py: 93/93 pass (dedupe preserves both bridges)
- tests/test_bitwarden_secrets.py + env_loader tests: 49/49 pass
V4A patch '*** Update File:', '*** Add File:', '*** Delete File:' headers
come from patch CONTENT, not the explicit `path=` argument. That makes
them attacker-influenceable through skill content, web extract output,
prompt injection, and other surfaces the agent processes. Headers like
'*** Update File: ../../../etc/shadow' would resolve relative to the
agent's cwd; in deployment configurations where that cwd is deep enough
to land outside Hermes' protected paths, the write could land somewhere
the agent operator did not intend.
Reject any V4A header containing a '..' path component before applying
the patch. The explicit `path=` argument on patch_tool is UNCHANGED —
the agent legitimately uses '..' there (e.g. `patch path='../other_module/x.py'`
from a worktree dir is normal cross-module editing).
Regression tests: V4A Update header with traversal rejected, V4A Add
header with traversal rejected, patch_v4a never invoked when rejection
fires.
Salvaged from PR #29395 by @waefrebeorn. The original PR added
has_traversal_component as a blanket reject on read_file_tool,
write_file_tool, patch_tool's explicit path, and search_tool — that
would break legitimate agent operation where '..' is normal. Also
dropped the over-eager skills_guard pattern additions
(pickle.loads/marshal.loads/ctypes.CDLL/importlib at high/critical
severity would false-positive on legit data-science and FFI skills).
Co-authored-by: teknium1 <127238744+teknium1@users.noreply.github.com>
