fix: move pairing_store and hooks init into GatewayRunner.__init__ - #115
Closed
Ridwannurudeen wants to merge 1 commit into
Closed
fix: move pairing_store and hooks init into GatewayRunner.__init__#115Ridwannurudeen wants to merge 1 commit into
Ridwannurudeen wants to merge 1 commit into
Conversation
…set to __init__ PairingStore and HookRegistry were initialised inside _flush_memories_before_reset(), which only runs on session expiry. start() calls self.hooks.discover_and_load() unconditionally, so GatewayRunner crashed with AttributeError on every fresh start before any session had expired. Move the two init blocks to the end of __init__() where they belong, and remove the duplicate from _flush_memories_before_reset(). Fixes NousResearch#110
Contributor
|
Fixed with fec5d59 Thank you |
jarvis-stark-ops
added a commit
to 1Team-Engineering/hermes-agent
that referenced
this pull request
Jun 16, 2026
* feat(mcp): scaffold hermes-tools-mcp stdio server with initialize NousResearch#109 Minimal stdio JSON-RPC server that responds to MCP 'initialize' with the expected protocol version + capabilities. Tools surface comes in follow-up commits. Env-bound to (profile, project) scope. Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com> * feat(mcp): tools/list returns filtered Hermes tool surface NousResearch#109 Filters the live Hermes tool registry to the v1 surface defined in spec §4.1. Includes special hermes_set_task tool for per-turn task context updates. Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com> * feat(mcp): tools/call dispatches to Hermes tool registry NousResearch#109 hermes_set_task updates per-turn task context (task_id, workspace). All other tools/call requests filter through ALLOWED_TOOLS, then dispatch via registry.dispatch() which handles both sync and async tools. Denied tools (kanban_dispatch etc.) return isError=true. Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com> * feat(kanban): add task_sessions table for relay scope tracking NousResearch#113 Per spec §4.7 — tracks (profile, project) → tmux session, claude session_id, FIFO path, MCP config path, project root, scope cwd, task count, timestamps. Helpers: upsert_session, get_session, touch_session, mark_compacted, evict_session, list_idle_sessions. * feat(relay): project derivation + task header builder NousResearch#114 derive_project maps a kanban workspace to its 'project' key: - scratch workspaces -> 'scratch' - git workspaces -> basename of git root - else raises ProviderError (B11: no silent default) build_task_header prepends '[task_id=... workspace=...]' to each user prompt so claude knows the current task without an MCP roundtrip. * feat(provider): register claude-code-relay ProviderProfile NousResearch#112 Registers the provider via the existing bundled-plugin pattern under plugins/model-providers/claude-code-relay/__init__.py. api_mode= 'claude_code_relay' is a new tag for the transport defined in the next commit. auth_type='none' because the claude binary handles its own OAuth — the relay just shells out. * feat(transport): scope spawn + send turn helpers NousResearch#112 ScopeContext bundles (profile, project, workspace). ensure_scope either finds a live tmux or invokes relay-spawn-scope.sh (writing the MCP config first). send_turn prepends the task header and relays through relay-send.sh, returning the captured response. Also added monkeypatch for derive_project_root in spawn test since tmp_path is not a git repo. * feat(transport): OpenAI-compat ChatCompletion response shape NousResearch#112 chat_completion() returns the standard OpenAI ChatCompletion object. Only the latest user message is forwarded — prior conversation rides on claude's in-session memory. Usage fields left null; /compact hook in next task captures /usage and backfills. * feat(kanban): /compact between-task hook for relay scopes NousResearch#115 After a task with provider=claude-code-relay completes or blocks, send /compact to its (profile, project) scope. Per B10, failures are logged but never block the completion. Records last_compacted_at on task_sessions. Also: add tasks.provider column to SCHEMA_SQL + additive ALTER migration; set conn.row_factory=sqlite3.Row in init_db test-fixture shortcut so complete_task works correctly in :memory: tests. Spike 2 (2026-06-15) confirmed /compact returns 'Not enough messages to compact' on short sessions; that's handled cleanly via the timeout in relay-send.sh slash-command extension. Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com> * feat(relay): idle-eviction cron step NousResearch#116 evict_idle_scopes deletes task_sessions rows whose last_used_at is older than HERMES_RELAY_IDLE_EVICT_SECS (default 86400 = 24h), and tears down the tmux session via relay-kill-scope.sh. Wire into nightly-hermes-maintenance follows separately. Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com> * feat(relay): boot-check that Hermes Stop hook is installed NousResearch#109 B3: Stop hook is mandatory for the relay to detect turn-end via FIFO. Boot-check raises ProviderError with installer path if the hook is missing — fail loud instead of silently hanging 180s per turn. Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com> * fix(relay): validate ScopeContext.profile + project against B8 regex Per code review on PR #20 — the shell relay-spawn-scope.sh validates ^[a-zA-Z0-9_-]+$ but Python-side _write_mcp_config writes /tmp/hermes-mcp-{slug}.json BEFORE invoking the shell. Slugs containing / or .. would have allowed path traversal out of /tmp. Add same regex check in ScopeContext.__post_init__. Closes B8 (spec §3). Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com> * fix(relay): chmod 0o600 on /tmp MCP config files Defense-in-depth: prevent workspace paths in the MCP config from leaking to other users sharing /tmp. Per code review on PR #20. Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com> * fix(kanban): honor HERMES_RELAY_BIN in /compact hook for consistency Transport reads HERMES_RELAY_BIN; the kanban_db /compact hook should too. Per code review on PR #20. Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com> * fix(relay): preserve task_sessions row when relay-kill-scope.sh fails Per code review on PR #20 — previously evicted the DB row even on kill failure, orphaning the tmux session with no record to retry. Now check rc; only evict DB row on rc=0. Failed scopes get retried on next cron tick. Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com> * fix(relay): log WARNING when relay-send exits 4 (Stop-hook timeout) Per code review on PR #20 — exit 4 means partial response, but callers had no signal. Add a logger.warning so operators see the partial in logs. Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com> --------- Co-authored-by: Jarvis <jarvis@Kaipos-Mac-mini.local> Co-authored-by: Claude Sonnet 4.6 <noreply@anthropic.com>
Meraniya
pushed a commit
to Meraniya/hermes-agent
that referenced
this pull request
Aug 6, 2026
* docs: evaluate Mnemosyne as a memory provider Adds docs/mnemosyne.md covering mnemosyne-oss/mnemosyne @ v3.15.0 (mnemosyne-hermes 0.5.0) as a candidate tenth external memory provider. Verdict: opt-in plugin, pinned — not a default, and not a store for CRM/vault content. In its favour: 2285 tests pass locally at CI parity in 5m23s; local-first is real (core/embeddings.py::_is_api_model gates the OpenRouter route behind an explicit opt-in); Actions are SHA-pinned; MCP SSE refuses a non-loopback bind without a token; 326 commits in 90 days with a two-day fix turnaround on NousResearch#491; and the provider implements all four MemoryProvider abstract methods plus most optional hooks. Against it: - CI sets MNEMOSYNE_NO_EMBEDDINGS=1, so the hybrid vector path — the whole value proposition — is never exercised. Every green test runs degraded. - Degradation is silent (NousResearch#474, NousResearch#518); reproduced locally at 1/3 recall on trivially easy queries with the model unreachable. - ALL_TOOL_SCHEMAS exposes 40 tools / ~6.5k tokens per request against a README claiming 23, and mnemosyne_triple_end is advertised over MCP with no _TOOL_HANDLERS entry, so calling it raises ValueError. - Two divergent in-tree copies of the provider reached by different install routes (hermes_memory_provider/ 3778 lines vs mnemosyne_hermes/ 2882). - NousResearch#487 (gateway RSS jump) and NousResearch#537 (hermes update drops the provider) land on us directly. - Open correctness bugs NousResearch#507/NousResearch#506/NousResearch#524/NousResearch#525/NousResearch#523, and NousResearch#482 with 50 of 106 config keys silently ignored. - Benchmark claims do not survive their own README: BEAM Recall@10 = 20% at every scale, end-to-end 65.2% against Hindsight's 73.4% in their table. Governance noted: the CLA pre-authorises relicensing under any future license, bus factor is ~2, and SECURITY.md points at the retired tracker while describing a confidential-issue feature GitHub does not have. Pilot conditions are listed in the doc. Documentation only. Co-Authored-By: Claude Opus 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01NxqJBGm7soA7QefpRBSDAu * fix(nix): refresh stale npm lockfile hashes for tui and web The fetchNpmDeps hashes in nix/tui.nix and nix/web.nix no longer matched the committed ui-tui/package-lock.json and web/package-lock.json, so the `nix` job failed on every PR with "Nix build failed due to stale npm lockfile hash". Drift is pre-existing on main, not introduced here: both nix files and both lockfiles are byte-identical between main and this branch, so the same check fails on main. Fixed here because it blocks this PR and the fix is mechanical — the replacement hashes are the values nix itself reported in the job log. Co-Authored-By: Claude Opus 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01NxqJBGm7soA7QefpRBSDAu --------- Co-authored-by: Claude <noreply@anthropic.com>
mansonyu2000
pushed a commit
to mansonyu2000/hermes-agent
that referenced
this pull request
Aug 10, 2026
brainstorm: agent-coding-workflow Skill → 5-phase plan tdd: pending (MIM tests in todo) verified: - zentao task list --execution=28 → 13 tasks created (IDs 32-44) - zentao story NousResearch#115 created under product NousResearch#10 - ZENTAO_DB_HOST=192.168.3.23 required for MySQL connection Added: - Zentao: 13 MIM tasks (F1-F3 + security) under execution NousResearch#28 - Zentao: story NousResearch#115 (MIM聊天V1) under product NousResearch#10 - tasks/mim-chat-v1-plan.md + tasks/mim-chat-v1-todo.md (双向关联zentao_id) - bin/sync-to-zentao.py (解析todo.md → 更新zentao任务状态) - docs/templates/: spec/todo/acceptance 模板 - docs/guides/dev-setup.md: 开发环境搭建指南 - docs/governance/CONTRIBUTING-WinPeek.md: WinPeek贡献指南 - docs/{reference-library,test,code,deploy}/README.md: 目录占位 - .claude/commands/: /onboard /status /handoff (本地斜杠命令) Changed: - CLAUDE-module.md: 新增 🚀新会话启动引导段落 - tasks/README.md: MIM状态→🟢进行中, 新增Zentao追踪表 - docs/README.md: 新增工作流目录 + 模板引用 - .zentao/mapping.json: 新增MIM任务映射 Fixed: - tools/winpeek_tools.py: _handle_mim_update_profile IDOR — 要求password认证 (security review finding: unauthorized profile update) Co-Authored-By: Claude <noreply@anthropic.com>
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
2 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
Fixes #110.
PairingStoreandHookRegistrywere initialised inside_flush_memories_before_reset(), which only runs when a session expires. Sincestart()callsself.hooks.discover_and_load()unconditionally on every startup,GatewayRunnercrashed with:on every fresh gateway start before any session had ever expired.
Root cause
A recent refactor (hooks system added ~Feb 16, follow-up commits Feb 23–27) accidentally left the initialisation block inside the wrong method.
Fix
Move the two blocks to the end of
__init__()where they belong, and remove the duplicate from_flush_memories_before_reset(). No logic changes — purely moving 6 lines to the correct scope.Test plan
hermes gateway startcompletes withoutAttributeErrorhermes gateway statusshows the service active_flush_memories_before_reset) still works as before —self.hooksandself.pairing_storeare now available throughout the object lifetime