mcp: add func-workflow prompt for guided Function lifecycle

[thonmay]

Changes

Adds a func-workflow MCP prompt that guides agents through the full Function lifecycle: create, build, and deploy.

Why Option 2 (add a real prompt) over Option 1 (set HasPrompts: false):
Setting the flag to false would be a minimal fix but leaves the capability unused. A guided workflow prompt is directly useful to agents and aligns with the MCP server's purpose of enabling end-to-end agentic usage of Functions.

Prompt behavior:

• Name: func-workflow
• Optional language argument to tailor guidance (e.g. go, python, node)
• Returns step-by-step instructions referencing the real MCP tools (create, build, deploy) and resources (func://languages, func://templates, func://function)
• HasPrompts: true remains correct now that a prompt is registered

Tests added (3):

• TestPrompt_Listed — prompt appears in ListPrompts response
• TestPrompt_Get — prompt returns valid message with create/deploy content
• TestPrompt_GetWithLanguage — language argument is reflected in output

/kind enhancement

Fixes #3734

Add MCP `func-workflow` prompt for guided Function lifecycle (create, build, deploy)

[knative-prow]

[APPROVALNOTIFIER] This PR is NOT APPROVED

This pull-request has been approved by: thonmay
Once this PR has been reviewed and has the lgtm label, please assign matejvasek for approval. For more information see the Code Review Process.

The full list of commands accepted by this bot can be found here.

Details

Needs approval from an approver in each of these files:

• OWNERS

Approvers can indicate their approval by writing /approve in a comment
Approvers can cancel approval by writing /approve cancel in a comment

[knative-prow]

Hi @thonmay. Thanks for your PR.

I'm waiting for a knative member to verify that this patch is reasonable to test. If it is, they should reply with /ok-to-test on its own line. Until that is done, I will not automatically test new commits in this PR, but the usual testing commands by org members will still work.

Regular contributors should join the org to skip this step.

Once the patch is verified, the new status will be reflected by the ok-to-test label.

I understand the commands that are listed here.

Details

Instructions for interacting with me using PR comments are available here. If you have questions or suggestions related to my behavior, please file an issue against the kubernetes-sigs/prow repository.

[gauron99]

🦴 CAVEMAN REVIEWER HAVE QUESTIONS 🦴

Thank for contribution. Before review, answer these. No answer, no review.

1. Explain

• In your own words — what this change do and how it work? Not copy from issue, not copy from PR template. YOUR words.
• This PR include describe tool from mcp: add describe tool for Function inspection #3736 AND prompt. Why bundle both? These separate changes — why one PR?

2. Why

• Issue mcp: either register prompts or set HasPrompts to false #3734 say two options: set HasPrompts: false OR add real prompt. You pick Option 2. Why? What make static text prompt more useful than just fixing flag?
• Would agent actually use this prompt in real workflow? Or this just exist so flag not lie?
• What specific problem this solve that agent can't already do by calling tools directly?

3. Testing

• You run MCP server and test prompt end-to-end with real agent? Not CI — you, on your machine.
• Do you screenshot your work? Caveman like see proof it work.

4. Understanding

• Explain in own words how MCP prompts different from MCP tools. When agent use prompt vs tool?
• What break if we no merge this?

We welcome all contributors. But change must be understood by author and solve real problem. PRs that no demonstrate this — closed.

[thonmay]

1. Explain

Adds func-workflow prompt: guided create → build → deploy message template. Takes optional language arg, returns step-by-step instructions referencing the right tools and resources.

Also includes describe tool from #3736. You're right, separate changes. I'll rebase this to prompt-only, keep describe in #3736.

2. Why

Option 1 (set flag false) = fix the lie, add no value.
Option 2 (add prompt) = fix the lie, add value.

Agent sees 17 tools and several resources. Doesn't know the order, doesn't know func://languages should be checked before create, doesn't know registry is needed on first deploy. The prompt gives that orchestration knowledge in one shot.

MCP clients (Claude Desktop, Cursor) expose prompts to users. User picks "func-workflow" → agent gets a structured plan. Tools are individual actions; prompt is the recipe.

Without it, agent relies on instructions.md (long, may get truncated in some clients).

3. Testing

Ran locally:

• prompts/list → shows func-workflow with [language] argument
• prompts/get(language="go") → returns full workflow text with "Use language: go" interpolated

Screenshots attached.

4. Understanding

Tools = model-controlled, agent calls them to DO things.
Prompts = user-controlled, inject structured guidance into conversation.

Agent uses tool when it knows what to do. User invokes prompt when they want guided help.

No merge = HasPrompts: true remains a lie. No workflow guidance for new users.

[thonmay]

[lkingland]

/ok-to-test

[codecov]

Codecov Report

✅ All modified and coverable lines are covered by tests.
✅ Project coverage is 57.10%. Comparing base (584c11e) to head (f6c8a57).
⚠️ Report is 1 commits behind head on main.

Additional details and impacted files

@@            Coverage Diff             @@
##             main    #3737      +/-   ##
==========================================
+ Coverage   57.03%   57.10%   +0.07%     
==========================================
  Files         182      183       +1     
  Lines       21376    21409      +33     
==========================================
+ Hits        12191    12225      +34     
  Misses       7953     7953              
+ Partials     1232     1231       -1     

Flag	Coverage Δ
e2e	35.75% <0.00%> (-0.07%)	⬇️
e2e go	31.36% <0.00%> (-0.06%)	⬇️
e2e node	27.20% <0.00%> (-0.05%)	⬇️
e2e python	31.73% <0.00%> (-0.06%)	⬇️
e2e quarkus	27.32% <0.00%> (-0.05%)	⬇️
e2e rust	26.74% <0.00%> (-0.04%)	⬇️
e2e springboot	25.27% <0.00%> (-0.05%)	⬇️
e2e typescript	27.31% <0.00%> (-0.05%)	⬇️
e2e-config-ci	28.11% <0.00%> (-0.05%)	⬇️
integration	17.14% <0.00%> (-0.04%)	⬇️
unit macos-14	45.17% <100.00%> (+0.08%)	⬆️
unit macos-latest	45.17% <100.00%> (+0.08%)	⬆️
unit ubuntu-24.04-arm	45.42% <100.00%> (+0.08%)	⬆️
unit ubuntu-latest	46.12% <100.00%> (+0.08%)	⬆️
unit windows-latest	45.21% <100.00%> (+0.07%)	⬆️

Flags with carried forward coverage won't be shown. Click here to find out more.

☔ View full report in Codecov by Sentry.
📢 Have feedback on the report? Share it here.

🚀 New features to boost your workflow:

• ❄️ Test Analytics: Detect flaky tests, report on failures, and find test suite problems.
• 📦 JS Bundle Analysis: Save yourself from yourself by tracking and limiting bundle sizes in JS merges.