stocktake 2026-11-11: 36 stale pages re-verified (0 stale); fallback chain rewritten 4/6 slots; CT333 'clawtest' naming claim corrected; nous 404 credits signature replaces refresh-token 401; estate shrank (VM 310 + CT601 gone)
- freshness: 36 stale / 93 fresh / 3 no-date -> 0 stale / 129 fresh / 3 no-date of 132 (same-day report regenerated after the corrections, per pattern 13) - skills: 299 skills / 67 categories (count unchanged; generator arithmetic re-verified by hand against the live tree: 300 SKILL.md files - 1 nested duplicate = 299) - runbooks/test-openrouter: live fallback chain rewritten (primary back to smart-route/custom:omni-101; mimo-v2.5, stepfun, qwen3.5/ollama-cloud out; guacano + kilo-reflection-beam-free in); ollama.com dropped from the URL test list; guacano/litellm noted as internal + key-gated - comparisons/hermes-vs-nanobot: the 2026-10-20 'CT333 is zeroclaw in Proxmox naming' claim was itself wrong - pct list and hostname both return clawtest - incidents/2026-10-22-nous: refresh-token failure gone from logs; new signature 404 insufficient_credits_for_paid_model from auxiliary.pulse_memory (ninth slot, back on nous, paid model) - open item revised - incidents/2026-08-29-gmail: email crons now ok but root cause unfixed 74 days (himalaya message read still in fetch_newsletters.py; goimapnotify back up PID 441; monitor still paused) - 30 further pages re-verified live and de-flagged with dated probe notes (ct207 ingress, CT205 guanaco v0.8.15, CT270 nextcloud 33.0.2.2, headless stack, hermes-ui both hosts, systemd units + cgroup limits, terminal/tool_search/mnemosyne/custom_providers config blocks, ops profile still deleted, hosts/pve guest lists) - log.md addendum + index.md date bump; no page deleted, no secrets written
This commit is contained in:
@@ -2455,3 +2455,15 @@ ledger activity remains the 2026-11-05T12:36:17 closures), so **nothing alerted*
|
||||
- No secrets, keys, tokens or passwords written.
|
||||
- **Post-audit housekeeping (01:0x AWST).** Vault sync committed in `/home/hermes/wiki` (**`3a19147`**, both vault conventions; 20 files, then amended with the freshness report). `wiki-maintain.sh` re-run **after** the page-date corrections → **skills 299 / 67 categories** (`systems/skills-index.md`), freshness **2026-11-10: 0 stale / 127 fresh / 3 no-date of 132**, Quartz rebuilt (807 files) and `hermes-wiki-serve.service` restarted. Two long-stale pages re-verified and de-flagged: [[systems/push-notifications]] (bridge sha256 `59cabfcc…` unchanged; `vory-push` relay **live 01:00:34, embedded in another Hermes process** — a `pgrep` on the script name is *expected* to return nothing while healthy) and [[incidents/2026-10-18-degenerate-stream-canned-substitute]] (6 runs since 10-18, **0 failed**; newest 2026-11-09 12:00:41 `completed`). **Pushed:** `6de35a3` (maintenance) → this.
|
||||
|
||||
### 2026-11-11 — Stocktake: 36 stale pages re-verified; 0 stale; one materially stale fallback chain rewritten
|
||||
|
||||
- **Freshness:** the 03:00 `wiki-maintain.sh` run flagged **36 stale / 93 fresh / 3 no-date of 132** — the whole 2026-10-20 cohort (runbooks, concepts, decisions, comparisons, experiments, seven systems pages, the Gmail + Mnemosyne-Capture incidents, `CONTRIBUTING.md`, `agents/index.md`). After the corrections below, the **same-day report was regenerated** (script → fix → re-run) and reads **0 stale / 129 fresh / 3 no-date of 132**. Quartz rebuilt (807 files) and `hermes-wiki-serve.service` restarted.
|
||||
- **Skills:** `systems/skills-index.md` regenerates at **299 skills / 67 categories** — unchanged count from 2026-11-10; the only diff is the date bump. Verified the generator's arithmetic against the live tree by hand: **300 `SKILL.md` files** (excluding `.archive/`), **1 nested duplicate** (`productivity/obsidian-git-sync/obsidian-git-sync/SKILL.md` duplicating the same category+name) → **299 unique (category, skill) pairs**. The count is correct; do not "fix" it to 300 or 303.
|
||||
- ⚠️ **Materially stale page rewritten — [[runbooks/test-openrouter]].** The 2026-10-20 fallback chain was wrong in four of six slots. Live `hermes fallback list` (verbatim recorded on the page): primary **`smart-route` via `custom:omni-101`**, then `meituan/longcat-2.5-preview:free` (nous), `deepseek-v4.1-flash` (opencode-zen), `deepseek-v4.1-flash` (custom:guacano), `kilo-reflection-beam-free` (custom:litellm), `poolside/laguna-m.1:free` (kilocode), `or-inclusionai_ling-2_6-1t` (custom). `mimo-v2.5`, `stepfun/step-3.7-flash:free` and `qwen3.5:397b` via ollama-cloud have **left** the chain; two new entries **joined** at 3 and 4. `ollama.com` dropped out of the runbook's URL test list; `guacano`/`litellm` are noted as **internal + key-gated (401 without a key)**. → [[concepts/model-routing]], [[comparisons/model-routers]]
|
||||
- ⚠️ **A false claim corrected — [[comparisons/hermes-vs-nanobot]].** The 2026-10-20 note claimed "CT333 is `zeroclaw` in the current Proxmox naming". Live `pct list` **and** `pct exec 333 -- hostname` both return **`clawtest`** — the wiki's original label was right; `zeroclaw` is only the daemon running inside it. The page now says so, dated.
|
||||
- ⚠️ **Incident signature change — [[incidents/2026-10-22-nous-portal-refresh-token-invalid]].** The refresh-token failure is **gone** (zero `invalid_grant` lines in `errors.log` through 2026-11-10; `auth.json` live, no `NOUS_API_KEY` in `.env`). What replaced it: **404 `insufficient_credits_for_paid_model`** for `anthropic/claude-haiku-4.5`, 3× on 2026-11-10. Cause identified: a **ninth auxiliary slot** `auxiliary.pulse_memory` points back at **`nous`** with a paid model — the 2026-10-25 remediation moved the original eight slots to `custom:omni-101` but this slot was added afterwards and is the only active nous path left. `nous` is still first in `fallback_providers`, but on a *free* model not observed failing. Open item revised from "re-authenticate" to "move `pulse_memory` or add credits".
|
||||
- **[[incidents/2026-08-29-gmail-read-state]] — the cron error state cleared, the root cause did not.** `goimapnotify` is **still active** (now PID 441), the Important-mail monitor (`d40085da631c`) is **still paused**, and `fetch_newsletters.py` **still calls `himalaya message read --preview`** on Anthony's INBOX — 74 days unfixed. The three email crons that read `error` on 2026-10-20 (`85edf6772bca`, `08eb86f02d37`, `06a0a1c8704f`) are now `ok` with `failure_streak: 0`, but green cron status is not evidence of safe behaviour: a job that marks mail read *and succeeds* still pollutes the mailbox. Both original mitigations are now inert.
|
||||
- **Live-vs-page diffs run (per skill pattern 12a), all claims held:** ct207 `cloudflared` ingress re-read in full (four Hermes entries unchanged, `mcp-portal` still absent — [[systems/cloudflare-tunnel]], [[infrastructure/domains-and-tunnels]]); CT205 Guanaco at `v0.8.15`/`5df435e` with Ollama 200 ([[systems/guanaco]]); CT270 `occ status` 33.0.2.2 installed/not-in-maintenance ([[systems/nextcloud]]); headless stack process/port state ([[systems/headless-server]], [[runbooks/restart-browser]]); hermes-ui 5174 on both CT460 and CT450 ([[systems/hermes-ui]]); all five `hermes-*` user units active + gateway/WebUI/Camofox effective cgroup limits ([[runbooks/restart-hermes]], [[runbooks/gateway-resource-resilience]]); `tool_loop_guardrails` and the whole `terminal:`/`tools:`/`memory.mnemosyne`/`custom_providers` blocks byte-identical; `ops` profile **still deleted 2026-10-11** while `multiplex_profiles: true` and both `profile_routes` still point at it — **31 days unresolved** ([[runbooks/multiplexer-setup]]); `/home/hermes/wiki` still a local-only repo, `/home/hermes/wiki-backups/` still absent, `~/.hermes/backups/` incident dirs intact; `m93p` **not** re-probed this pass (its 2026-10-20 guest list stands as written).
|
||||
- **Estate shrank again (probed, not inferred):** `qm list` on `pve` now holds only VMs **100, 101, 110, 282** (VM **310** gone; the wiki had listed it) and `pct list` holds 17 CTs with **CT 601 gone**. VM 403/VM 401 and the `open-claw` Tailscale peer remain absent on both hypervisors → [[agents/open-claw]], [[agents/index]], `CONTRIBUTING.md`.
|
||||
- **Delivery verified at the destination, not assumed:** before the commit, `http://100.118.5.51:9120/` and the changed pages (incl. `runbooks/test-openrouter`, `concepts/model-routing`, `incidents/2026-08-29-gmail-read-state`) all returned **200**, as did the external route `https://hermes.kangaroo-eel.ts.net/wiki/`. `CONTRIBUTING.md`'s access table was re-verified the same way and its `agents/open-claw/` scope row struck through ([[CONTRIBUTING]]).
|
||||
- **No page deleted, no decision reversed, no secrets written.** No credential value appears in any page (the Gitea remote's embedded token is deliberately not reproduced). One pre-existing broken wikilink left as-is: `systems/guanaco.md` → `[[systems/ollama]]` (present in HEAD, page already qualifies it "if exists").
|
||||
|
||||
Reference in New Issue
Block a user