[aj-emerich]

Changes

New page: claude-code-quickstart/index.mdx

• Sidebar: "Claude Code Quickstart", right after Quickstart.
• It's a tutorial. The Develop with AI pages (Docs MCP server, Agent Skills) and the kestractl page stay the reference, and the quickstart links to them instead of repeating their skills list or setup options.
• Includes a troubleshooting table and a note that the example flow's hourly trigger keeps running until you disable it.

CodeTabs.vue

• A tab can set its own lang, so a group can mix a shell command with a JSON config. Cursor's MCP config uses this.
• Groups that share a storageKey now switch together when the reader clicks a tab. Before, they only matched on the next page load.
• The migration tutorial is the only other page using the component. Its rendered tabs are byte-identical before and after the change. The only difference there is that its three kestractl / curl groups now switch together.

kestractl page fixes

• The install script now installs kestractl 3.x, which works with Kestra 2.x and 1.3. The page said it installed 2.0 and that it only worked with Kestra 2.0. The source is the install script's own comments.
• Documents where the script installs the binary and how to choose the folder with INSTALL_DIR, and links the releases page for Windows and other platforms.

Cross-links and nav

• getNavigationTree.ts: adds the page to Get Started.
• Quickstart and Develop with AI: link to the new page.
• llms.txt: adds an entry under "AI and agents".

Testing

• Claude Code, end to end agaie:
  • Installed the skills with npx skills add.
  • Prompt 1 wrote `github_stars.
  • Prompt 2 validated, deployed, and ran the flow to SUCCESS, and reported the star count
    with kestractl v3.6.2.
  • Test flows and executions were deleted afterwards.
• Cursor and Codex: verifiedent cursor|codex installs to.agents/skills/`.
  • The Cursor MCP JSON matches .
  • The Codex MCP commands come from OpenAI's Codex docs.
  • Neither agent was run end to
• CodeTabs: checked in headless Chrome on the dev server:
  • Both pages hydrate and highls.
  • Tab selection syncs across groups and persists after reload.
  • The Cursor JSON tab is highl
• Unit tests: navigation tests pass (17/17).

[github-actions]

🔎 OpenGrep ✅ (0 findings)

Found 0 finding(s).

🐷 TruffleHog ✅ (0 secrets)

No secret found.

☁️ Cloudflare Worker Preview Deployed!

🔗 https://ks-docs-claude-code-quickstart-docs.kestra-io.workers.dev
🔗 https://653a7ec3-docs.kestra-io.workers.dev

🔦 Lighthouse Benchmark

Tested on 2026-10-30 20:07 UTC  ·  links point to https://ks-docs-claude-code-quickstart-docs.kestra-io.workers.dev
Compared against main baseline from 2026-10-30
Runner CPU index per shard: 2392, 2305, 2499, 2483 (baseline: 2479, deltas hidden on 5 page(s))

Scores (0–100, higher is better)

Page	Performance	Accessibility	Best Practices	SEO
Home	87	85	77	92
Get Started	81	84	77	100
Pricing	97	92	77	100
Enterprise	95	84	77	100
Cloud	92	81	77	100
Features	96	92	77	92
About Us	81	92	77	100
Security	94	94	77	100
Docs Landing	89	88	77	92
Contribute to Kestra (simple docs)	98	88	77	92
Flow (full featured docs)	96	89	77	92
Blog Index	98	91	77	100
Blog Post (sample)	97	88	77	100
VS Page (sample)	97	89	77	100
Customers Landing	95	89	77	100
Customer Story (sample)	99	91	77	100
Plugins Landing	97	81	77	92
Plugin Page (sample)	98	88	77	100
Plugin Debug Page (sample)	99	88	77	100
Plugin Debug Return Page (sample)	94	88	77	100
Blueprints Landing	87	91	77	100
Blueprint Audit Logs CSV Export	88	92	77	100

Core Web Vitals (lower is better)

Page	LCP	FCP	TBT	CLS	Speed Index
Home	1.41 s	0.42 s	210 ms 🔻	0.000	1.53 s
Get Started	1.56 s	0.45 s	304 ms	0.000	1.13 s
Pricing	0.41 s	0.41 s	134 ms	0.000	1.03 s
Enterprise	1.35 s	0.41 s	100 ms	0.003	0.84 s
Cloud	1.69 s	0.42 s	95 ms	0.049	1.00 s
Features	1.06 s	0.44 s	116 ms	0.000	1.01 s
About Us	0.41 s	0.41 s	428 ms	0.000	0.90 s
Security	1.26 s	0.36 s	136 ms	0.000	1.08 s
Docs Landing	1.09 s 🔻	0.38 s	243 ms	0.001 🟢	0.97 s
Contribute to Kestra (simple docs)	0.95 s	0.38 s	79 ms	0.000	0.94 s
Flow (full featured docs)	0.61 s 🟢	0.61 s 🔻	142 ms	0.000	1.21 s
Blog Index	0.99 s	0.38 s	66 ms	0.000	1.16 s 🔻
Blog Post (sample)	1.14 s	0.37 s	80 ms	0.000	0.75 s
VS Page (sample)	0.45 s	0.45 s	143 ms	0.005	0.74 s
Customers Landing	1.35 s	0.48 s	62 ms	0.000 🟢	1.48 s
Customer Story (sample)	0.48 s	0.48 s	71 ms 🟢	0.000	1.09 s 🔻
Plugins Landing	1.03 s	0.43 s	117 ms 🟢	0.000	1.10 s
Plugin Page (sample)	0.93 s	0.47 s	94 ms	0.000	1.00 s
Plugin Debug Page (sample)	0.94 s	0.41 s	56 ms	0.000	0.83 s
Plugin Debug Return Page (sample)	1.01 s	0.42 s	142 ms	0.000	1.54 s
Blueprints Landing	2.27 s	0.56 s	98 ms 🔻	0.000 🟢	1.09 s
Blueprint Audit Logs CSV Export	0.54 s	0.54 s	235 ms	0.000	1.86 s

Legend

🟢 improved  ·  🔻 regressed  ·  (blank) no significant change
Score threshold: ±10 pts  ·  Metric threshold: ±30% of baseline

Median of repeated runs: Home x5, Get Started x3, Pricing x3, Enterprise x3, Cloud x3, Features x3, About Us x3, Security x3, Docs Landing x5, Contribute to Kestra (simple docs) x3, Flow (full featured docs) x3, Blog Index x3, Blog Post (sample) x3, VS Page (sample) x3, Customers Landing x3, Customer Story (sample) x3, Plugins Landing x3, Plugin Page (sample) x3, Plugin Debug Page (sample) x3, Plugin Debug Return Page (sample) x3, Blueprints Landing x5, Blueprint Audit Logs CSV Export x5. A single run of these swings 20+ points between runners.
The sample is measured across 4 parallel runners, so each page carries its own CPU index (Lighthouse's benchmarkIndex). Lighthouse does not normalise for host CPU, so a page's deltas are hidden when its runner differs from the baseline's by more than 10%.
A page's deltas are also hidden when the baseline measured it a different number of times, since comparing a median against a single run mostly reports the baseline's spread. That resolves once main carries the same run counts.

View full Lighthouse HTML report for a page

Full per-page Lighthouse Results (LHR) are attached as the lhr-reports-shard-* artifacts on this run. Download and unzip one, then open https://googlechrome.github.io/lighthouse/viewer/ and drop the <page>-lhr.json file into the page to see every audit, opportunity, and diagnostic.

Open this run's artifacts

[wrussell1999]

Sweet!