7.6 KiB
title, type, status, created, updated, verified_on, review_after, confidence, tags, sources
| title | type | status | created | updated | verified_on | review_after | confidence | tags | sources | ||||||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Conduit | system | active | 2026-10-24 | 2026-10-25 | 2026-10-25 | 2026-11-08 | high |
|
|
Conduit
Purpose
Hermes on Anthony's phone. A dashboard plugin + iOS app pair that pushes cards, prompts,
location and health data between the Hermes gateway and the iPhone, and lets a voice call
be started from the phone where the agent can speak and listen (conduit-voice sessions).
Related: systems/mnemosyne (memory backend the voice session reads), systems/hermes-ui
(the browser desktop), systems/push-notifications (the older hermes-push bridge).
Where it lives
- Plugin tree:
/home/hermes/.hermes/plugins/conduit_push - Version:
0.69.0(plugin.yamlanddashboard/manifest.json), git HEADf000bf0— “Pulse: list candidates, decide in code, journal them (#337)”. 22 commits on from the0.64.0/eaa14berecorded at the 2026-10-24 audit (upstream#324–#338: Pulse cards/decisions/journal, a Live Activity messaging tool, contact cards, a self-update gateway restart,channels.yaml, App Store builds following the archived plugin version). - Upstream:
https://github.com/kaishi00/hermes-conduit-notifier.git(origin, fetch+push) - ⚠ The working tree still carries the uncommitted local patch (
M dashboard/plugin_api.py, the memory-timeout fix in incidents/2026-10-22-conduit-memory-start-timeout) — now five releases behind HEAD (0.64.0→0.69.0), so the chance a plugin update resets the tree and drops it is rising, not falling. - Skills shipped with the plugin (4, under
plugins/conduit_push/skills/):calling-the-user,reading-health-data,using-calendar,using-location. - Data stores (
~/.hermes/):conduit-health.sqlite3,conduit-location.sqlite3,conduit-phone.sqlite3, plus JSON state —conduit-push.json(credential / gateway id / installation id / relay url / e2e — no values recorded here),conduit-calls.json,conduit-settings.json,conduit-present.json,conduit-desktop-views.json.
Current state (verified 2026-10-25 00:2x–00:5x AWST)
- Dashboard gateway: listening on
0.0.0.0:9119; process started 2026-10-23 18:54:43 and not restarted since — so it is still the process that picked up the memory-timeout patch, and the0.64.0→0.69.0upstream commits are not live. - Device: one registered — iPhone (
id 70aa63f3…,kid c24e8215…), app version2.0.0, featureschat-cards,info-cards,live-cards,maps-navigate,media-cards,show-cards; tzAustralia/Perth; sharing window 2026-10-16 → 2027-04-22. Paired 2026-10-22 12:54:39.last_seen/last_sent/last_contactall read 2026-10-23 19:10:30 AWST (~29 h stale at this audit) while health rows kept arriving until 2026-10-24 09:59:59 and window rollups were recomputed at 13:56:35 — so those three fields track phone-status contact, not data flow. (The 10-24 page recorded 19:02:16 from the same store; the epoch reads 19:10:35.) - Call settings:
enabled,when_asked,decides,alertsall true; rate limitsmin_gap_s 120/per_hour 6/per_day 20;avoid_unknown_placefalse. - Recent calls — none.
conduit-calls.json → historyis empty: no call has been recorded since the memory-timeout fix landed, so that fix still has no positive evidence behind it. The “two call sessions closed cleanly 00:03:23 / 00:06:57” claim previously on this page was a false positive — thoseendsentries are the two cron sessions of that night, andendsrecords any turn-end that fires a watch, not calls → incidents/2026-10-24-conduit-call-request-refused. - Data volume (2026-10-25 00:2x):
conduit-health.sqlite3569.9 MB (+8.2 MB in 25.7 h), holding 1,520,229events(1,510,929sample/ 6,070sleep/ 3,230workout), 3,028bucketsand 124windowsfor its single device; newest row 2026-10-24 09:59:59, newestbuckets.date2026-10-24;conduit-location.sqlite380 KB;conduit-phone.sqlite3284 KB. ⚠ The 10-24 page recorded 308buckets/ 62windows; a directCOUNT(*)reads 3,028 / 124, so the earlier pair was a narrower grouping — quote these. No retention or compaction has been observed — the store grows from ingestion alone and is worth watching against the CT460 disk budget.
Calls from Hermes (the conduit_call_user tool)
The plugin registers a conduit_call_user tool (toolset conduit, 📞) for every session and
ships the matching skill calling-the-user. Hermes exposes it through lazy tool discovery
(tool_search), so it is not in a session's headline tool list — it has to be looked up.
Gate conditions (call_tool.py): a real session (subagents are refused), a paired device,
settings.enabled, and either when_asked (with the ask recorded as user-originated) or decides;
the call rings when the asking turn ends and plays the reason first. Limits: min_gap_s 120 /
6 per hour / 20 per day.
⚠ Known failure, 2026-10-24: three requests to be phoned were refused by the agent, which asserted the capability did not exist without calling the tool or loading the skill → incidents/2026-10-24-conduit-call-request-refused.
Memory integration
The voice session (conduit-voice) reaches Mnemosyne through the plugin's own memory layer, not
through Mnemosyne's system_prompt_block() — recall is wrapped in a per-request deadline
(MEMORY_TIMEOUT_S, default 60 s as of 2026-10-23, overridable with
CONDUIT_MEMORY_TIMEOUT_S; it was a hard-coded 10 s before that, which was too short for
Mnemosyne's 10–45 s cold start). If the provider fails to start, the plugin logs a warning and
proceeds without recall rather than abandoning the call.
See incidents/2026-10-22-conduit-memory-start-timeout and systems/mnemosyne.
Maintenance commands
# what version is installed / is the tree patched?
git -C ~/.hermes/plugins/conduit_push log --oneline -1 && \
git -C ~/.hermes/plugins/conduit_push status --short
# is the gateway up, and since when?
ss -tlnp | grep 9119
# did a voice call lose its memory context?
grep -h "Conduit" ~/.hermes/logs/agent.log | grep -i "memory" | tail
# did a real call happen? (history = calls; ends = turn-ends, NOT calls)
python3 -c "import json,os;d=json.load(open(os.path.expanduser('~/.hermes/conduit-calls.json')));print(len(d['history']),'calls;',len(d['ends']),'turn-ends')"
# state files (no secrets in the wiki)
ls -la ~/.hermes/conduit-*.json ~/.hermes/conduit-*.sqlite3
Open items
- Commit or upstream the
plugin_api.pytimeout patch so a plugin update cannot revert it — it is now five releases behind HEAD. - Verify a call end-to-end. No call has been placed since the fix, so
MEMORY_TIMEOUT_S = 60is still unexercised by a real call, and the 10-24 refusal means the chat route has never worked → incidents/2026-10-24-conduit-call-request-refused. - No retention policy on
conduit-health.sqlite3(569.9 MB and growing). - The Mnemosyne cold start that motivated the patch is still slow (10–45 s); only the budget was raised. → systems/mnemosyne