fix: prevent premature subagent termination, add attack planning framework, and fix TUI lag#702
Open
pennywise-01 wants to merge 510 commits into
Open
fix: prevent premature subagent termination, add attack planning framework, and fix TUI lag#702pennywise-01 wants to merge 510 commits into
pennywise-01 wants to merge 510 commits into
Conversation
Updates LiteLLM from ~1.80.7 to ~1.81.1 which includes full support for z.ai (Zhipu AI) provider using the 'zai/model-name' format. This enables Strix to work with z.ai subscription credentials by setting STRIX_LLM="zai/glm-4.7" with appropriate LLM_API_KEY and LLM_API_BASE environment variables. Changes: - Updated litellm version constraint in pyproject.toml - No breaking changes to Strix API or configuration Closes #ISSUE_ID (to be linked if applicable) Signed-off-by: legendevent <legendevent@users.noreply.github.com>
The memory compressor was calling litellm.completion() without passing the api_key and api_base parameters, causing authentication errors when LLM_API_KEY is set but provider-specific env vars (OPENAI_API_KEY, etc.) are not. This matches the pattern used in dedupe.py.
Bumps [pypdf](https://github.com/py-pdf/pypdf) from 6.6.0 to 6.6.2. - [Release notes](https://github.com/py-pdf/pypdf/releases) - [Changelog](https://github.com/py-pdf/pypdf/blob/main/CHANGELOG.md) - [Commits](py-pdf/pypdf@6.6.0...6.6.2) --- updated-dependencies: - dependency-name: pypdf dependency-version: 6.6.2 dependency-type: indirect ... Signed-off-by: dependabot[bot] <support@github.com>
Remove hardcoded github.com/gitlab.com/bitbucket.org host lists from infer_target_type. Instead, detect git repositories on any host by querying the standard /info/refs?service=git-upload-pack endpoint. Works for any self-hosted git instance.
Improve the finish_scan tool schema to produce more professional pentest reports: expand parameter descriptions with structural guidance, rewrite recommendations example with proper urgency tiers instead of Priority 0/1/2, fix duplicated section titles, and clean up informal language.
Enable native text selection across tool components and agent messages with automatic clipboard copy, toast notification, and decorative icon stripping. Replace Padding wrappers with Text to support selection across multiple renderables.
Bumps [pillow](https://github.com/python-pillow/Pillow) from 11.3.0 to 12.1.1. - [Release notes](https://github.com/python-pillow/Pillow/releases) - [Changelog](https://github.com/python-pillow/Pillow/blob/main/CHANGES.rst) - [Commits](python-pillow/Pillow@11.3.0...12.1.1) --- updated-dependencies: - dependency-name: pillow dependency-version: 12.1.1 dependency-type: indirect ... Signed-off-by: dependabot[bot] <support@github.com>
Bumps [cryptography](https://github.com/pyca/cryptography) from 44.0.1 to 46.0.5. - [Changelog](https://github.com/pyca/cryptography/blob/main/CHANGELOG.rst) - [Commits](pyca/cryptography@44.0.1...46.0.5) --- updated-dependencies: - dependency-name: cryptography dependency-version: 46.0.5 dependency-type: indirect ... Signed-off-by: dependabot[bot] <support@github.com>
Bumps [protobuf](https://github.com/protocolbuffers/protobuf) from 6.33.4 to 6.33.5. - [Release notes](https://github.com/protocolbuffers/protobuf/releases) - [Commits](https://github.com/protocolbuffers/protobuf/commits) --- updated-dependencies: - dependency-name: protobuf dependency-version: 6.33.5 dependency-type: indirect ... Signed-off-by: dependabot[bot] <support@github.com>
… and CVSS Replace 12 flat parameters (code_file, code_before, code_after, code_diff, and 8 CVSS fields) with structured nested XML fields: code_locations with co-located fix_before/fix_after per location, cvss_breakdown, and cwe. This enables multi-file vulnerability locations, per-location fixes with precise line numbers, data flow representation (source/sink), CWE classification, and compatibility with GitHub/GitLab PR review APIs.
- Specify encoding="utf-8" in registry.py _load_xml_schema() - Specify encoding="utf-8" in skills/__init__.py load_skills() - Prevents cp949/shift_jis/cp1252 decoding errors on non-English Windows
… multi-part suggestions Rewrote the code_locations parameter description to make fix_before/fix_after semantics explicit: they are literal block-level replacements mapped directly to GitHub/GitLab PR suggestion blocks. Added guidance for multi-part fixes (separate locations for non-contiguous changes like imports + code), common mistakes to avoid, and updated all examples to demonstrate multi-line ranges.
* feat: add to readme new keys * feat: shoutout strix models, docs * fix: mypy error * fix: base api * docs: update quickstart and models * fixes: changes to docs uniform api_key variable naming * test: git commit hook * nevermind it was nothing * docs: Update default model to claude-sonnet-4.6 and improve Strix Router docs - Replace gpt-5 and opus-4.6 defaults with claude-sonnet-4.6 across all docs and code - Rewrite Strix Router (models.mdx) page with clearer structure and messaging - Add Strix Router as recommended option in overview.mdx and quickstart prerequisites - Update stale Claude 4.5 references to 4.6 in anthropic.mdx, openrouter.mdx, bug_report.md - Fix install.sh links to point to models.strix.ai and correct docs URLs - Update error message examples in main.py to use claude-sonnet-4-6 --------- Co-authored-by: 0xallam <ahmed39652003@gmail.com>
OpenAI's Responses API rejects reasoning.effort on non-reasoning models like gpt-4o with `unsupported_parameter`, so any scan with the default STRIX_REASONING_EFFORT=high against gpt-4o crashed at the first model call. drop_params=True absorbs the rejected param on LiteLLM-routed models but the SDK's native OpenAI path has no equivalent. Lift model_supports_reasoning to a public helper that strips litellm/, any-llm/, openai/ prefixes and falls back to last-segment lookup so prefixed forms like anthropic/claude-opus-4-7 resolve through the bare model_cost entry. make_model_settings regains model_name and skips Reasoning() when the registry doesn't confirm support. uses_chat_completions_tool_schema reuses the same helper (was duplicating the lookup under a misleading name).
Register a litellm.success_callback that captures kwargs['response_cost'] into a new observed-cost bucket on LLMUsageLedger. record() skips the tokens-times-registry estimate for LiteLLM-routed models so we do not double-count with the callback; OpenAI direct routes keep estimating since LiteLLM is not invoked for them. Per-agent attribution for LiteLLM-routed calls is apportioned by token share at to_record() time.
…sestrix#562) Co-authored-by: 0xallam <ahmed39652003@gmail.com>
* fix: resolve pre-commit check failures - Change RuntimeError to TypeError for type validation in report/writer.py - Update pyupgrade to v3.21.2 for Python 3.14 compatibility * feat(cli): add --max-budget-usd flag Raises BudgetExceededError in ReportUsageHooks after each LLM call when accumulated cost reaches the limit, with clean "stopped" status and child-agent cancellation in non-interactive mode. * test: add budget enforcement unit tests 7 tests covering no-budget, under-budget, at-limit, over-limit, error message content, None report state, and exception hierarchy. Also adds pytest/pytest-asyncio to dev deps and a mypy override for tests. * fix(budget): validate positive budget and check the live cost ledger Two hardening fixes for --max-budget-usd enforcement: - Reject non-positive budgets. ReportUsageHooks now raises ValueError for max_budget_usd <= 0, and the CLI validates the flag via a custom argparse type so '--max-budget-usd 0' fails fast with a friendly message instead of silently killing the scan on the first model response. - Read the live cost. The budget check now reads ReportState.get_total_llm_cost() (the live ledger) instead of the persisted run-record snapshot, so it stays accurate even when a usage save fails after a model call. * fix(budget): stop the entire scan deterministically when the limit is hit Previously a BudgetExceededError was handled per-agent: it was swallowed in interactive mode (the loop kept waiting), a child's error escaped its detached task as an unretrieved-exception warning, the parent was never released from wait_for_message, and the stop was logged at ERROR with a traceback as if the agent had failed. Replace that with a single scan-wide signal on the coordinator: - AgentCoordinator.trigger_budget_stop() sets a flag and wakes every parked agent; wait_for_message returns as soon as the flag is set. - The run loops check coordinator.budget_stopped and raise to exit cleanly, marking themselves 'stopped'. The root's exception reaches run_strix_scan's handler, which cancels descendants and tears the scan down once; child exceptions are swallowed in their detached task. - The budget stop is logged at INFO, not as a failure. This is deterministic regardless of tree depth or which agent first sees the limit, fixing the interactive/TUI hang where a deep agent's stop never reached a parked root. Also re-raises BudgetExceededError explicitly in the stream handler so it can't be mistaken for the LiteLLM 'after shutdown' race. * fix(budget): treat a budget stop as a clean stop in the TUI Add an explicit BudgetExceededError handler in the TUI scan thread so that, if the error ever reaches it, the budget stop is logged as a graceful stop rather than surfaced as a red scan error by the broad 'except Exception'. The runner normally absorbs the error and returns cleanly, so this is defensive depth for a money-spending feature. * docs(cli): document --max-budget-usd behavior and limitations Clarify that the budget is cumulative across all agents, checked after each model response, that the scan stops cleanly (not as a failure), that the value must be > 0, and that spend can slightly overshoot due to in-flight calls and best-effort cost estimation. * Apply suggestions from code review Co-authored-by: greptile-apps[bot] <165735046+greptile-apps[bot]@users.noreply.github.com> --------- Co-authored-by: greptile-apps[bot] <165735046+greptile-apps[bot]@users.noreply.github.com>
* fix: resolve pre-commit check failures - Change RuntimeError to TypeError for type validation in report/writer.py - Update pyupgrade to v3.21.2 for Python 3.14 compatibility * chore: add pytest test infrastructure Mirror the layout introduced on feature/438-token_budget: pytest + pytest-asyncio dev deps, asyncio_mode auto, a tests.* mypy override, and pytest in the mypy pre-commit hook deps so the tests/ package type-checks. * feat: add --mount and large-target pre-flight for local repos (usestrix#492) Large local targets were copied into the sandbox file-by-file via the SDK LocalDir entry, which stalls on big repos and could leave /workspace empty. - --mount <path> bind-mounts a host directory read-only at /workspace/<subdir> instead of copying it, bypassing the per-file stream. - A size pre-flight (STRIX_MAX_LOCAL_COPY_MB, default 1024) fails fast with a clear message suggesting --mount when a non-mounted local target is too big. * fix: reject empty --mount paths An empty or whitespace-only --mount value resolves to the current working directory and would silently bind-mount it into the sandbox. Reject it. * fix: dedupe local targets so a dir is never both copied and mounted If the same directory is passed via --target and --mount (or as duplicate values), it previously produced two targets — copied AND bind-mounted, and the copied one could trip the size pre-flight. Dedupe by resolved path, preferring the bind mount. * fix: treat non-positive STRIX_MAX_LOCAL_COPY_MB as disabled Previously a value of 0 (or negative) made every local target count as oversized, aborting all local scans. Now <= 0 disables the pre-flight. * fix: log unreadable subtrees during size pre-flight os.walk silently swallowed directory-listing errors, so a permission-denied subtree could make a large repo under-count and slip past the pre-flight. Surface such omissions via an onerror warning. * docs: document --mount and STRIX_MAX_LOCAL_COPY_MB Add CLI reference + example for --mount, document the size pre-flight env var, note the read-only-is-not-a-hard-boundary caveat and that remote repos are not size-checked, and clarify the backends docstring on when bind mounts apply. * Update strix/interface/main.py * Update strix/runtime/docker_client.py ---------
…trix#596) Co-authored-by: Claude Opus 4.8 <noreply@anthropic.com>
…sestrix#637) Line 72 was over-indented, causing an IndentationError on import of strix/report/writer.py and breaking main. Also bump the mirrors-mypy pre-commit hook to v1.17.1 to avoid the mypy 1.16.0 internal crash (python/mypy#19412) on openai/_client.py.
Co-authored-by: Alex Schapiro <46074070+bearsyankees@users.noreply.github.com>
* fix: capture cost for streamed LiteLLM responses * docs: note LiteLLM streaming metadata callbacks
…ation, Django (usestrix#617) * Add five community security skills for agent specialization Expand coverage with OAuth flow testing, AWS misconfigurations, prototype pollution, insecure deserialization, and Django framework playbooks. * Address Greptile review feedback on AWS and deserialization skills - Use head-bucket for S3 existence checks instead of duplicating s3 ls - Add Node.js to insecure_deserialization frontmatter description * Clarify S3 existence vs public listing checks in aws skill Split unauthenticated enumeration into separate head-bucket/HTTP and s3 ls steps with interpretation guidance per review. * some tools ads --------- Co-authored-by: bearsyankees <bearsyankees@gmail.com>
…sestrix#626) * feat(report): SARIF 2.1.0 emitter for CI / code-scanning integration Strix emits CSV + markdown + JSON but no SARIF, so findings can't feed GitHub code-scanning, an ASPM, or any SARIF-consuming CI gate. Add a stdlib-only emitter (strix/report/sarif.py) and always write findings.sarif from ReportState._save_artifacts, beside the existing artifacts. Design invariants (learned from running this in production): - Stable partialFingerprints.primaryLocationLineHash per finding, so a re-scan that re-words a title doesn't churn code-scanning alert IDs. - Class/category hashing so the same vuln class maps to a stable ruleId across scans rather than drifting. - Findings with no code location anchor to SECURITY.md with a synthetic location marker instead of being silently dropped. - Always emit (even with zero findings) so a clean re-scan overwrites a stale findings.sarif and code-scanning auto-resolves fixed alerts. - tool.driver.version reports the strix package version. - Fully isolated in its own try/except: a SARIF build error must never break the CSV/MD/run-record path. Verified end-to-end on v1.0.4 against a SQLi/cmd-inj/weak-hash fixture: 3 findings -> valid SARIF 2.1.0, 3 results, real code locations, distinct per-finding fingerprints. Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com> * fix(report): complete SARIF code scanning metadata --------- Co-authored-by: bearsyankees <bearsyankees@gmail.com>
…ework, and fix TUI lag Subagent stop bug: - Remove stop_agent as suggested option in finish_scan docstring and rejection message - Rewrite root_agent.md Completion section to require waiting for all agents - Add warning to stop_agent docstring against using it to clear active agents Attack planning: - Add Attack Planning section to root_agent.md with think + load_skill framework - Add ATTACK PLANNING METHODOLOGY to system_prompt.jinja vulnerability_focus - Replace brute-force decomposition with technology-aware agent spawning TUI lag: - Batch SDK events via 150ms flush timer instead of per-event call_from_thread - Index events by agent_id for O(1) lookup instead of O(n) list scan - Cap events per agent at 500 with automatic pruning of oldest events - Window event rendering to last 200 events instead of full re-render - Reduce animation timer from 60ms to 150ms
Contributor
Greptile SummaryThis PR addresses three distinct bugs: premature subagent termination (by tightening
Confidence Score: 3/5The guidance-only and docstring changes are safe to ship; the TUI batching rework has a threading model error that can silently drop display events and calls a non-thread-safe Textual API from a background thread. Four of the six changed files are prompt templates and docstrings with no runtime risk. The TUI changes in app.py introduce a flush callback that always routes through the RuntimeError exception path (because call_from_thread is invoked on the main thread rather than a background thread), and a race window where events appended to the buffer between the snapshot copy and the clear are permanently lost from the display. Additionally, set_timer is called from the background scan thread without thread-safety guarantees. These issues affect only the TUI display layer, not scan results, but the batching mechanism as written doesn't behave as intended. strix/interface/tui/app.py — the event-batching and flush-timer implementation needs a second look for correct threading; strix/interface/tui/live_view.py — the O(n) self.events.remove() maintenance is worth cleaning up. Important Files Changed
Prompt To Fix All With AIFix the following 3 code review issues. Work through them one at a time, proposing concise fixes.
---
### Issue 1 of 3
strix/interface/tui/app.py:1421-1431
**Flush timer always takes the RuntimeError fallback path**
`_flush_event_buffer` is invoked by a Textual timer, which means it runs on the Textual main thread. Calling `call_from_thread` from the main thread always raises `RuntimeError`, so `_record_sdk_events_batch` is dead code — every flush goes through the per-event fallback loop. Separately, there is a race window between `batch = self._event_buffer[:]` (snapshot) and `self._event_buffer.clear()`: the background SDK thread can append an event into the buffer after the snapshot is taken but before `clear()` executes. That event is removed from the buffer without being in `batch`, silently dropping it from the TUI display.
### Issue 2 of 3
strix/interface/tui/app.py:1416-1419
**`set_timer` called from a background thread**
`_capture_sdk_event` is passed as the `event_sink` callback and is invoked from the background asyncio scan loop, not the Textual main thread. `set_timer` is not documented as thread-safe in Textual; the supported path for scheduling work from a background thread is `call_from_thread`. Also, the `if self._event_flush_timer is None` guard and the subsequent assignment are not atomic, so two rapid callback invocations from the scan loop can both observe `None` and register duplicate flush timers.
### Issue 3 of 3
strix/interface/tui/live_view.py:286-293
**`self.events.remove(d)` is O(n) overhead on an otherwise-unused list**
After the migration to `_events_by_agent`, `self.events` is no longer queried anywhere — `events_for_agent` and `has_events_for_agent` both use the new index. The per-prune `list.remove()` call is an O(n) linear scan across the entire global event list just to keep `self.events` in sync with a structure nothing reads. For long scans with many active agents the list can grow to `MAX_EVENTS_PER_AGENT × num_agents` entries, making each pruning step increasingly expensive.
Reviews (1): Last reviewed commit: "fix: prevent premature subagent terminat..." | Re-trigger Greptile |
![greptile-apps[bot]](https://avatars.githubusercontent.com/in/867647?s=60&v=4){kind=small}
Comment on lines
+1421
to
+1431
| def _flush_event_buffer(self) -> None: | ||
| self._event_flush_timer = None | ||
| if not self._event_buffer: | ||
| return | ||
| batch = self._event_buffer[:] | ||
| self._event_buffer.clear() | ||
| try: | ||
| self.call_from_thread(self._record_sdk_event, agent_id, event) | ||
| self.call_from_thread(self._record_sdk_events_batch, batch) | ||
| except RuntimeError: | ||
| self._record_sdk_event(agent_id, event) | ||
| for agent_id, event in batch: | ||
| self._record_sdk_event(agent_id, event) |
Contributor
There was a problem hiding this comment.
_flush_event_buffer is invoked by a Textual timer, which means it runs on the Textual main thread. Calling call_from_thread from the main thread always raises RuntimeError, so _record_sdk_events_batch is dead code — every flush goes through the per-event fallback loop. Separately, there is a race window between batch = self._event_buffer[:] (snapshot) and self._event_buffer.clear(): the background SDK thread can append an event into the buffer after the snapshot is taken but before clear() executes. That event is removed from the buffer without being in batch, silently dropping it from the TUI display.
Prompt To Fix With AI
This is a comment left during a code review.
Path: strix/interface/tui/app.py
Line: 1421-1431
Comment:
**Flush timer always takes the RuntimeError fallback path**
`_flush_event_buffer` is invoked by a Textual timer, which means it runs on the Textual main thread. Calling `call_from_thread` from the main thread always raises `RuntimeError`, so `_record_sdk_events_batch` is dead code — every flush goes through the per-event fallback loop. Separately, there is a race window between `batch = self._event_buffer[:]` (snapshot) and `self._event_buffer.clear()`: the background SDK thread can append an event into the buffer after the snapshot is taken but before `clear()` executes. That event is removed from the buffer without being in `batch`, silently dropping it from the TUI display.
How can I resolve this? If you propose a fix, please make it concise.
Comment on lines
1416
to
+1419
| def _capture_sdk_event(self, agent_id: str, event: Any) -> None: | ||
| self._event_buffer.append((agent_id, event)) | ||
| if self._event_flush_timer is None: | ||
| self._event_flush_timer = self.set_timer(0.15, self._flush_event_buffer) |
Contributor
There was a problem hiding this comment.
set_timer called from a background thread
_capture_sdk_event is passed as the event_sink callback and is invoked from the background asyncio scan loop, not the Textual main thread. set_timer is not documented as thread-safe in Textual; the supported path for scheduling work from a background thread is call_from_thread. Also, the if self._event_flush_timer is None guard and the subsequent assignment are not atomic, so two rapid callback invocations from the scan loop can both observe None and register duplicate flush timers.
Prompt To Fix With AI
This is a comment left during a code review.
Path: strix/interface/tui/app.py
Line: 1416-1419
Comment:
**`set_timer` called from a background thread**
`_capture_sdk_event` is passed as the `event_sink` callback and is invoked from the background asyncio scan loop, not the Textual main thread. `set_timer` is not documented as thread-safe in Textual; the supported path for scheduling work from a background thread is `call_from_thread`. Also, the `if self._event_flush_timer is None` guard and the subsequent assignment are not atomic, so two rapid callback invocations from the scan loop can both observe `None` and register duplicate flush timers.
How can I resolve this? If you propose a fix, please make it concise.
Comment on lines
+286
to
+293
| if len(agent_events) > MAX_EVENTS_PER_AGENT: | ||
| dropped = agent_events[:len(agent_events) - MAX_EVENTS_PER_AGENT] | ||
| for d in dropped: | ||
| try: | ||
| self.events.remove(d) | ||
| except ValueError: | ||
| pass | ||
| del agent_events[:len(agent_events) - MAX_EVENTS_PER_AGENT] |
Contributor
There was a problem hiding this comment.
self.events.remove(d) is O(n) overhead on an otherwise-unused list
After the migration to _events_by_agent, self.events is no longer queried anywhere — events_for_agent and has_events_for_agent both use the new index. The per-prune list.remove() call is an O(n) linear scan across the entire global event list just to keep self.events in sync with a structure nothing reads. For long scans with many active agents the list can grow to MAX_EVENTS_PER_AGENT × num_agents entries, making each pruning step increasingly expensive.
Prompt To Fix With AI
This is a comment left during a code review.
Path: strix/interface/tui/live_view.py
Line: 286-293
Comment:
**`self.events.remove(d)` is O(n) overhead on an otherwise-unused list**
After the migration to `_events_by_agent`, `self.events` is no longer queried anywhere — `events_for_agent` and `has_events_for_agent` both use the new index. The per-prune `list.remove()` call is an O(n) linear scan across the entire global event list just to keep `self.events` in sync with a structure nothing reads. For long scans with many active agents the list can grow to `MAX_EVENTS_PER_AGENT × num_agents` entries, making each pruning step increasingly expensive.
How can I resolve this? If you propose a fix, please make it concise.Note: If this suggestion doesn't match your team's coding style, reply to this and let me know. I'll remember it for next time!
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
20 participants
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Summary
This PR fixes three bugs identified through systematic debugging:
Root agent prematurely stops subagents — The root agent would call stop_agent on active children to bypass the active-agent check in finish_scan, orphaning their work and producing incomplete reports.
Agent doesn't think deeply about attack planning — The root agent received high-level vulnerability categories but no structured thinking framework. It blindly spawned agents for every vulnerability type instead of reasoning about what's relevant to the target.
TUI freezes during multi-agent scans — The TUI became unresponsive for 1-2 minutes when multiple subagents were running, due to event flooding, unbounded event lists, and aggressive timers.
Changes
1. Prevent Premature Subagent Termination
tools/finish/tool.py
completed/crashedare safeskills/coordination/root_agent.md
tools/agents_graph/tools.py
2. Attack Planning Framework
skills/coordination/root_agent.md
load_skillbefore planningthinkto reason through target technology, attack surface, technique selection, defense analysis, bypass strategies, chaining opportunities, and priority orderingagents/prompts/system_prompt.jinja
<vulnerability_focus>with 5-step process: load skill → think → match techniques to technology → plan bypasses → chain for impact3. TUI Lag Fixes
interface/tui/app.py
call_from_threadper event (~50/sec → ~6/sec message queue pressure)interface/tui/live_view.py
dict[str, list[dict]]for O(1) lookup instead of O(n) list scanMAX_EVENTS_PER_AGENT = 500with automatic pruning of oldest events