Researches our key material suppliers and world commodity prices on the 1st and 15th, and files a short brief that also lands in the company knowledge base — rebuilding the intelligence habit we lost when the old agent was retired.
A short brief on the 1st and 15th: what changed, why it matters to us, and a source per claim. The brief files itself into the company wiki inbox, so the knowledge brain learns it too.
The old supplier-intelligence job was retired in May with its prompts lost. Rebuilt fresh — and the first brief immediately earned its keep: it caught a 15-year-high wool price (≈2× year-over-year), a key supplier expanding into our region, and debunked a stale certification claim against the official register.
Twice-monthly rhythm
Suppliers, prices, certifications, capacity news
Primary sources; registers over rumours
What changed + why it matters + sources
Into the wiki inbox → the company brain learns it
For implementers and the technically curious. The full build sheet — verified queries, thresholds, and build notes — lives on the specs page.
| System | Role in this loop |
|---|---|
| Web researchwith source discipline | Every claim carries a source; official registers outrank news posts. |
| Company wiki + GBrainknowledge base | The brief drops into the wiki's inbox folder; the overnight router ingests it, so the company AI brain can answer from it. |
| Hermesagent runtime on our server | The scheduler that wakes the loop up. Each loop is a cron job under a Hermes profile; the planned bizops profile will host the business digests (IT loops run under vpsops). |
| Email renderer + gwsrender_email.py | All digests pass through one shared renderer: Huxberry-branded HTML, tables for repeated rows, a coral 'needs your response' box when the loop has questions, and an arrow link on every record. Sent from the loops mailbox via the Google Workspace CLI. |
| Agent Vaultcredential store | Where API keys live. Loops get read-only credentials from the vault; nothing is hardcoded in scripts. |
| Model / brain | What it does here |
|---|---|
| GPT-5.5 via Codexthe bulk-work model | Wrote and maintains the mechanical parts — SQL, diffing, digest assembly. Effectively free on our existing subscription, so routine cycles cost almost nothing. |
| Claude Opusthe judgment model | Reviews alert wording, thresholds, and anything a human will read and act on. Post-Fable, Opus owns everything that ships. |
The watchlist file (suppliers, topics, standing questions) plus prior briefs in the wiki — each new brief builds on what's already known.
Email brief + wiki ingestion. Telegram summary optional later.
Read-only research; a human-reviewable brief is the only output. Anything entering the knowledge base is Opus-reviewed first.