Claus Ibsen created CAMEL-26700:
-----------------------------------
Summary: camel-jbang - an About screen in the Overview: what an
integration does, at a glance and without AI
Key: CAMEL-26700
URL: https://issues.apache.org/jira/browse/CAMEL-26700
Project: Camel
Issue Type: New Feature
Components: camel-jbang
Reporter: Claus Ibsen
Fix For: 4.24.0
When the terminal UI is opened on an integration you have not seen before, no
screen gives the gist of what it does. The Overview is an operations table with
one row and a chart. The Diagram takes a while to read, so people go and look
at the source. The AI summary (CAMEL-25143) needs an AI, takes time, and is a
lot of text with few visual aids.
h3. Proposal
The Overview gets two sub-tabs: *About* and *Integrations* (today's table and
chart). About is pre-selected when one integration runs, or when a project is
opened and not running. Integrations is pre-selected when several run, and
Enter opens About for the selected one.
About shows, at once and with no AI:
* *Header*: runtime, files, number of routes, uptime, state; and a purpose line
taken from, in this order: the comment at the top of the main source file, the
route descriptions, a saved AI summary (marked ✦, with its age), or a hint on
where to add one. The screen never calls an AI itself.
* *In → Routes → Out*: what starts the integration (entry points, grouped by
kind, with rates); the routes, with the internal ones ({{direct:}}, {{seda:}})
indented under the route that calls them; and the systems it talks to. The
internal routes are shown, as they help to understand it; they fold only when
there are more than N routes and not enough room.
* *Traits*: error handling, notable EIPs, data formats, beans, REST and HTTP
servers, chaos.
* *Config / Needs*: properties, secrets, and what it takes to start (infra
services, properties without a value).
* *Watch*: live errors, health, circuit breakers, the dead letter channel, the
findings of the source analysis, and the last result of the CVE Audit tab with
its age. About never runs the audit itself, as it sends the dependency
coordinates to OSV.dev.
* *Next*: keys that lead deeper. Enter on a route opens it on the Diagram,
Enter on an endpoint opens the Endpoint tab, {{2}} opens the source, {{!}} is
chaos here, {{?}} asks the AI with About as context.
It is built from {{ProjectOverview.analyze}} (entry points, systems, routes,
links, findings), the route structure and the live status, so it also works
from the sources alone for a project that is not running.
Design sketch: https://claude.ai/artifact/2H927wodRp3FAy1BbDbw3P
h3. Stages (one PR each)
# The Overview sub-tabs, with About pre-selected for one integration, built
from the sources: header, purpose line, In, Routes, Out, Traits.
# The live part: rates and sparklines, chaos and circuit-breaker marks, Watch
(errors, health, dead letter channel, CVE).
# The keys that lead deeper, folding for big integrations, the narrow layouts.
# "Good to know" hints: what happens when it fails, what is dropped, what does
not wait, what a restart forgets, twice or out of order, what is exposed or
insecure. These need a curated rule set kept outside the terminal UI (see
below), so this is a follow-up.
# The same as text: {{camel get about}}, and the live part in the
{{camel_project_overview}} MCP tool, from one model class in camel-jbang-core.
The hints should not be a set picked while building the screen. They belong to
a curated set of rules that more tools share: today's {{ProjectOverview}}
findings, the {{camel_security_scan}} and harden tools of camel-jbang-mcp, the
Doctor, and the option metadata of the catalog ({{secret}}, {{security =
"insecure:*"}}). That is for a separate JIRA.
_Claude Code on behalf of Claus Ibsen_
--
This message was sent by Atlassian Jira
(v10.3.26#10030026)
