docs: overhaul init/config docs + screenshots for netclaw 0.24.0 (epic #55)#69
Merged
Merged
Conversation
…erence Redirect channel, search, webhook, exposure, skills, and telemetry pages from the old netclaw init wizard steps to the re-entrant netclaw config dashboard, document canonical channel-ID resolution, and add a new cli/config.md page wired into the sidebar nav. Homepage and installation copy updated for the 5-step init flow. Screenshot re-capture and the init.md/quickstart.md rewrites are tracked separately under epic #55.
Distill the netclaw.dev house voice into a reusable netclaw-docs skill (voice rules, punctuation policy, per-page-type quick reference, terminology, pre-ship checklist) and add an AGENTS.md entry point that points at both CLAUDE.md (the authoring process) and the skill (how pages should read). CLAUDE.md now links the skill from its workflow. The skill is force-added past the .claude/ gitignore so the voice contract is shared, not local-only.
Full rewrite of cli/init.md and getting-started/quickstart.md to the shipped 5-step wizard (Provider -> Identity -> Security Posture -> Enabled Features -> Health Check; Personal skips features for a 4-step flow; health check auto-launches chat). Correct the cli/config.md dashboard area list and status strings, narrow the channel-ID migration callouts to AllowedChannelIds (Default channel fields are still resolved at runtime), fix the SearXNG probe as a non-blocking warning, drop the stale 'Mattermost not in status' caveat, and correct the skill-sources auto-detection, webhook validation, and alert-type claims against 0.24.0 source. Bump documentedVersion to 0.24.0. Screenshots still reference the pre-0.24.0 captures and are re-captured next.
Re-capture the init wizard (5-step flow) and the netclaw config Settings Areas dashboard against the released 0.24.0 binary in a throwaway Docker container, replacing the stale 11-step captures. Wire the new shots into cli/init.md, quickstart.md, and cli/config.md, and delete 26 obsolete init-* images. Capturing against the real binary surfaced a doc fix: the provider list ships seven entries (Anthropic, GitHub Copilot, Ollama, OpenAI, llama.cpp / vLLM, OpenRouter, Venice.ai), not the six the dev-branch source listed.
Link validation (pr_validation) regressed because the GitHub repo was transferred from Aaronontheweb/netclaw to netclaw-dev/netclaw, and the old paths now 404: - Changelog release-note bodies (fetched at build time in lib/releases.js) hard- code Aaronontheweb/netclaw issue/PR links; rewrite them to the current repo so the ~280 changelog links resolve again. - The footer Apache-2.0 link and the SPEC-001 reference pointed at blob/main, but the default branch is dev (no main); repoint to blob/dev. Cuts broken-link instances from 452 to 22 (the remainder is pre-existing rot on pages outside this change: dead third-party URLs and bot-blocked domains).
Repoint links that 404 after upstream restructures (surfaced by link-validation): - MCP spec: spec.modelcontextprotocol.io -> modelcontextprotocol.io/specification - llama.cpp server README: examples/server -> tools/server - Ollama modelfile: github docs -> docs.ollama.com/modelfile - NIST AI RMF: /artificial-intelligence/ -> /itl/ - Akka.NET persistence: persistence/overview.html -> event-sourcing.html
Capture the Channels, Search, Telemetry & Alerting, Inbound Webhooks, Exposure Mode, Skill Sources, Browser Automation, and Workspaces Directory editors against 0.24.0 in a seeded Docker container, and wire them into cli/config.md (replacing the screenshot TODO placeholders).
- Cloudflare Agent Skills Discovery RFC repo was renamed: cloudflare/agent-skills-spec -> cloudflare/agent-skills-discovery-rfc (4 pages). - ClawHub marketplace moved: clawhub.openclaw.ai -> hub.openclaw.ai.
A Sonnet QA pass and source verification corrected fabricated/stale specifics across the redesigned pages, each checked against 0.24.0 source: - Proactive send tool is the generic send_channel_message (channel_key + destination), not per-channel send_slack_message/send_discord_message/ send_mattermost_message; corrected on slack/discord/mattermost/webhooks/ reminders. Discord DOES support proactive messaging (DiscordProactiveOutboundClient) — fixed the stale 'no proactive tools' claim. Reverted a regression that dropped Discord from the valid current_session reminder origins (SetReminderTool allows Slack/Discord/ Tui/SignalR). - Removed the non-existent 'netclaw skill feed add/list' command (skill-feeds/overview/skill-server); remote feeds are config-TUI/JSON only. - Other verified corrections: real Slack debug-log strings, lowercase search Backend enum values, actual tunnel/auth startup error text, webhook retry delays (2s/4s) and DropOldest behavior, real built-in skill names, exact Skill Sources menu labels, PowerShell -Channel param, the config dashboard row list, and screenshot captions aligned to the real images. Build green. Deferred to the screenshot follow-up: Personal-posture 'Step X of 4' wizard shots and the 5-button approval prompt.
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
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.
Reworks the getting-started flow, CLI reference, channel/config/skills pages, and screenshots for the
netclaw init/netclaw configTUI redesign shipped in 0.24.0 (netclaw-dev/netclaw#1368). Delivers the copy/structure + screenshot scope of epic #55.What changed
cli/init.mdrewritten to the shipped 5-step wizard (Provider → Identity → Security Posture → Enabled Features → Health Check; Personal skips Features for a 4-step flow; the health check auto-launchesnetclaw chat).getting-started/quickstart.mdrewritten to the 5-step flow; channel/exposure steps moved out tonetclaw config.cli/config.md— the re-entrant Settings Areas dashboard (live status summaries, autosave-on-completion, resolve-before-add channels, multi-webhook telemetry, SearXNG reachability probe, routednetclaw provider/netclaw modelhandoffs).netclaw initsteps tonetclaw config, with the canonical channel-ID migration callout.netclaw-docsvoice/style skill +AGENTS.mdto keep future doc work consistent.documentedVersion→0.24.0.Corrections found vs. the planning forecast
Validated against 0.24.0 source/specs and the running binary, which caught several places the released product differs from the dev-branch source the epic was written against:
netclaw status— stale caveat removed.DefaultChannelNameis still resolved at runtime and not deprecated — breaking-change callout narrowed toAllowedChannelIds.netclaw configis a manual add flow (no Claude Code/Open Code auto-detect).Link-validation (CI) repair
pr_validationregressed because the GitHub repo was transferred (Aaronontheweb/netclaw→netclaw-dev/netclaw), which 404'd ~280 changelog links (fetched at build time bylib/releases.js), the footerblob/main/LICENSE(nomainbranch), and several moved third-party URLs. Fixed all of them: changelog body rewrite, footer/SPECmain→dev, MCP/llama.cpp/Ollama/NIST/getAkka URL-moves, the agent-skills RFC rename (agent-skills-spec→agent-skills-discovery-rfc), and the ClawHub move. CI is green.Verification
npm run buildpasses (137 pages); every referenced screenshot exists in the source-of-truth dir; no broken image refs.pr_validation(Build Astro + Link and URL Validation) is green.Remaining screenshot follow-ups (non-blocking)
A few secondary captures are left as honest
<!-- TODO(screenshots) -->, not fake images: the init existing-install action menu + reset-scope chooser, the config channels resolve-in-action sub-flow, and the first-chat TUI. The capture rig (Docker + VHS) is in place to add these.Closes
Closes #55
Closes #56
Closes #57
Closes #58
Closes #59
Closes #60
Closes #61
Closes #62
Closes #63
Closes #64
Closes #65
Closes #66
Closes #67
Closes #68