Models
The support agent
Answers a customer from one product's own record, and is scanned on the way out for the things it must never say.
A product surface, not a published identifier. It runs on the shared inference path and carries no identifier of its own.
What it is for
The support agent answers questions from an operator's customers about that operator's product. Its whole context is the product record: the specifications, the offer row with its observed figures and their bases, and the store manifest.
It is the only FlowFinds surface whose output reaches someone who is not the account holder, which is why two of its rules are enforced on the text rather than asked for in its instructions.
What it can do
Record-bound answering
A question the record answers is answered from the record. A question it does not is refused by name — we do not have that on record, so we will not guess — and handed to a person.
Output scanning, not output trust
It never states a shipping cost, customs charge, duty, import fee or delivery caveat, and the outgoing text is scanned for them. A rule that lives only in an author's intention is not a control.
The tense rule carries through
An observed undercut may be stated flatly because it was observed. The margin is declared by the owner and is never quoted to a customer at all. A price carries the date it was checked, and any failed refresh, through the same resolver the offer screen and the storefront use, so the three cannot disagree.
The inference path
| Property | Value |
|---|---|
| Runtime | Codex CLI, one ephemeral session per request |
| Model | gpt-5.5 |
| Reasoning effort | medium |
| Sandbox | read-only by default; workspace-write only where an edit is being applied |
| Isolation | The caller supplies an already-authorised account workspace as the working directory. No session state is shared between founders, and workspace-write never extends above that directory. |
| Pinned flags | exec --ephemeral --ignore-user-config --skip-git-repo-check --output-last-message |
The session is ephemeral and ignores any user configuration on the host, so a response cannot be shaped by state left behind on the machine. A refusal from the runtime is read out of its own event stream and passed through in its own words: an account out of credits reports that, rather than a generic failure the operator can do nothing about.
How it is measured
Not covered by commerce-v1.
Because no recorded run exercises this surface, this page states no measurement of it at all. The benchmark and its system cards are published under research; they describe commerce-reasoning-v1 and nothing else.
Its limits
- No authority to issue a refund, change a price, or alter an order. It answers.
- Where the store has no returns or warranty terms on record, it says so rather than stating one.
- It cannot see anything outside the claims of the account it belongs to. An order identifier belonging to someone else reads exactly like one that does not exist.
- commerce-v1 does not exercise this surface.
Where these facts come from
Each line names a file in the implementation. The descriptions on this page are a transcription of those files and are marked provisional at their source until they have been checked against it again.
- Capabilities implemented today — flowfinds-ai-tool/PRODUCT-MAP.md
- The inference path — flowfinds-ai-tool/backend/codex_cli.py
The others
- commerce-reasoning-v1 — The decision engine an operator talks to: it investigates a question with application tools before it answers, and it names the evidence it used.
- The find engine — Scores a candidate product on three independently sourced legs, and refuses to average them into one comfortable number.
- The store editor — Turns an instruction in the operator's own words into a proposed change with a cited reason, a rejected alternative, and a reversibility flag.
- The organic research engine — Researches an audience per platform and produces a framing brief. It has no posting function, and no credential it could post with.
Back to all models, or read what the engine is used for under the product.