feat(tui): show advertised skill count for remote skill servers#1452
Merged
Aaronontheweb merged 2 commits intoJun 22, 2026
Merged
Conversation
The Skill Sources config screen now surfaces how many skills each remote server
advertises, so operators can see what a feed offers without leaving the TUI.
- The reachability probe parses the RFC agent-skills index it already fetches and
returns the skill count; it shows in the add / Test-connection feedback and, after
"Rescan all", as a "| N advertised" badge on the inventory row.
- "advertised", not "loaded": it is the count the server publishes. The daemon may
load fewer after content-scan / hash / minimum-version filtering, so the label is
honest about what the number measures.
- Rescan all refreshes counts off the loop (one probe per ENABLED remote server) via
a shared probe helper, caches them by source name, and posts a completion summary
("Updated skill counts for K of N skill server(s); J did not respond") instead of a
premature green success.
- The count cache is invalidated on every config reload (rename / remove / URL change /
disable), so a stale or misattributed count cannot survive a config change.
- A stored bearer token that will not decrypt is skipped (no anonymous probe) and
surfaced as a Rescan warning, never silently degraded to an unauthenticated request.
- Interactive add / test probes stay capped at 10s for snappy feedback; the background
Rescan refresh honors the feed's full configured timeout (matching the daemon).
Tests cover advertised-count parsing, the disabled / undecryptable-token / unreachable
paths, and cache invalidation on disable.
Aaronontheweb
enabled auto-merge (squash)
June 19, 2026 19:26
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
1 participant
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Summary
The Skill Sources config screen now shows how many skills each remote server advertises, so operators can see what a feed offers without leaving the TUI.
/.well-known/agent-skills/index.json) it already fetches and returns the count. It shows in the add / Test-connection feedback and, after Rescan all, as a| N advertisedbadge on the inventory row.minimumDaemonVersionfiltering — the label is honest about what it measures.Updated skill counts for K of N skill server(s); J did not respond) instead of a premature green success.This shipped with a
/code-review ultra-style pass folded in — the cache-staleness, disabled-feed, silent-fallback, network-failure-visibility, and reuse findings were all addressed before this PR.Validation
Netclaw.Cli.Tests: 1112/1112 (38 SkillSources tests — advertised-count parsing, disabled / undecryptable-token / unreachable paths, cache invalidation on disable)dotnet slopwatch analyze: no new violationsghcr.io/netclaw-dev/skillserver(seeded with 5 skills viaskillserver publish-all): a netclaw image built from this branch reads the RFC index and renders| 5 advertisedon the Rescan-all'd row.Smoke note
No native smoke tape: a render-level tape would need a fake skill-server stood up inside the tape (no existing tape stands up a data server, and there's no pre-tape setup hook). The count is a render-independent string covered by the VM integration tests, and the behavior is validated end-to-end against the real skillserver image above.