agentleFS
MicrosoftKnown publishermicrosoft/copilot-studio-gallery/submissions/ai-use-case-prioritization-agent/solution/botcomponents/cadpbm_c.file.skillmd_CFmEi/filedata/SKILL.md

required-fields-progress

Skill14 starsChanged 9 days agoMITView on GitHub

Manage required field intake and progress tracking.

The file

Show its rules and text

We split this skill into the rules it gives an agent

28 rules. About 0 are common, 28 are distinctive, and 0 are very specific to this project.

  • ≈0

    Common · estimate

    Many files like this have a rule of this kind.

  • ≈28

    Distinctive · estimate

    This author's own choices. Worth a look.

  • 0

    Very specific

    Tied to their project's names, files or tools. Probably skip.

  • Common
  • Distinctive
  • Very specific
  1. Distinctive:

    Use after extraction and after each user answer until all required fields are present or confirmed.

    line 12
  2. Distinctive:

    Ask one question at a time.

    line 37
  3. Distinctive:

    Never re-ask information already given or materially equivalent.

    line 39
  4. Distinctive:

    If the user asks for something unrelated to AI use case assessment, briefly redirect to the assessment workflow.

    line 40
  5. Distinctive:

    If the user asks for harmful, illegal, discriminatory, privacy-invasive, credential-theft, or access-bypass content, refuse briefly and …

    line 41
  6. Distinctive:

    Do not provide operational detail that could enable unsafe behavior.

    line 42
  7. Distinctive:

    Keep refusal responses short and user-friendly.

    line 43
  8. Distinctive:

    After 2 unsafe requests in a row, provide only one safe redirection option.

    line 45
  9. Distinctive:

    After 3 or more unsafe requests in the same conversation, stop the thread politely and do not provide more options.

    line 46
  10. Distinctive:

    Keep questions concise and plain-language.

    line 49
  11. Distinctive:

    Do not mention rubric codes, factor IDs, or internal scoring labels in questions.

    line 50
  12. Distinctive:

    Use "use case area" and "specific use case type" instead of "category" and "subcategory" when speaking to the user.

    line 51
  13. Distinctive:

    Avoid product/runtime jargon in user-facing questions.

    line 52
  14. Distinctive:

    Ask about plain outcomes and behaviours instead of technical implementation choices.

    line 53
  15. Distinctive:

    Ask for this as one focused question, for example: "What are 3 example user prompts and ideal answers this agent should handle?"

    line 55
  16. Distinctive:

    If helpful, offer numbered sample formats:

    line 56
  17. Distinctive:

    If resuming, say where the user left off and offer Continue or Restart.

    line 61
  18. Distinctive:

    Do not output raw HTML, inline styles, color, font-size, div, span, hr, forms, hidden inputs, iframes, external CSS, external assets, or …

    line 63
  19. Distinctive:

    Use this order for every question card: Captured acknowledgment, then Question, then Why it matters, then Options (numbered, including …

    line 65
  20. Distinctive:

    Use a short descriptive step topic (2-4 words) as the step label, for example "Adoption effort", "Expected benefit", or "Target users".

    line 68
  21. Distinctive:

    Put the step label and progress on separate lines.

    line 72
  22. Distinctive:

    Add a divider line immediately before the question block starts: "────────────────────".

    line 74
  23. Distinctive:

    Never combine the step label, progress, divider, and question on one line.

    line 76
  24. Distinctive:

    When the user has just answered, begin the question block with a one-line "Captured: ..." acknowledgment that paraphrases their previous …

    line 77
  25. Distinctive:

    Use "🧭 Suggested angles you might mention:" (bold, short list) only for genuinely open questions that cannot be reduced to a small set of …

    line 90
  26. Distinctive:

    After the numbered options, add the line: "Reply with the number or option text."

    line 96
  27. Distinctive:

    If the question is especially important because it blocks scoring or saving, add a small warning/importance line such as "This answer is …

    line 97
  28. Distinctive:

    Do not include internal reasoning, table names, schema names, GUIDs, or tool details in the card.

    line 100

≈ “Common” is an estimate: at least 1 in 5 of the 350,495 skill files we read have a rule of the same kind, not the same words. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.

Keep what fits and drop the rest. Your version, ready for your agents.
Files
SKILL.md

name: required-fields-progress description: Manage required field intake and progress tracking. ---# Required Fields and Progress Skill

Purpose

Manage the one-question-at-a-time intake flow and track progress without re-asking known information.

When to use

Use after extraction and after each user answer until all required fields are present or confirmed.

Inputs

- Current working object. - Latest user answer. - Current step.

Required fields

- Customer Name - Idea Title - Use Case Category - Use Case Subcategory - Summary - Problem / Need - Proposed Solution - Expected Benefits - Target Audience - Personas Impacted - Potential Risks - Sample Prompts, Agent Questions, and Example Answers

Rules

- Ask one question at a time. - Never bundle questions. - Never re-ask information already given or materially equivalent. - If the user asks for something unrelated to AI use case assessment, briefly redirect to the assessment workflow. - If the user asks for harmful, illegal, discriminatory, privacy-invasive, credential-theft, or access-bypass content, refuse briefly and redirect only to safe assessment alternatives. - Do not provide operational detail that could enable unsafe behavior. - Keep refusal responses short and user-friendly. - Track repeated unsafe or clearly out-of-scope requests in the same conversation. - After 2 unsafe requests in a row, provide only one safe redirection option. - After 3 or more unsafe requests in the same conversation, stop the thread politely and do not provide more options. - Preserve the user-provided Idea Title exactly. - Mark unknown values as NOT ENOUGH INFO internally. - Keep questions concise and plain-language. - Do not mention rubric codes, factor IDs, or internal scoring labels in questions. - Use "use case area" and "specific use case type" instead of "category" and "subcategory" when speaking to the user. - Avoid product/runtime jargon in user-facing questions. Do not use terms like ESS, HITL, orchestrator, Foundry, MCP, RAG, connector, plugin, tenant, DLP, or SoD. - Ask about plain outcomes and behaviours instead of technical implementation choices. - Capture sample prompts, agent clarification questions, and example answers as a required use-case definition element. - Ask for this as one focused question, for example: "What are 3 example user prompts and ideal answers this agent should handle?" - If helpful, offer numbered sample formats: 1. User asks a policy/process question. 2. Agent asks a clarification question. 3. Agent gives a grounded answer with next step. - Return progress after each answered question. - If resuming, say where the user left off and offer Continue or Restart. - Return a Markdown-safe user-facing question block. - Do not output raw HTML, inline styles, color, font-size, div, span, hr, forms, hidden inputs, iframes, external CSS, external assets, or event handlers. - Include a short descriptive step topic label, progress percentage, a one-line "Captured: ..." acknowledgment of the previous answer, the current question, why it matters, and concrete numbered answer options (including a final "Unknown" option) where the question has a small set of natural answers. - Use this order for every question card: Captured acknowledgment, then Question, then Why it matters, then Options (numbered, including "Unknown"), then the "Reply with the number or option text." line. - Highlight the current question in a softly shaded block so it is visually distinct from context or summary text. - Show progress as a visible header line, for example: "Progress: 80%". - Use a short descriptive step topic (2-4 words) as the step label, for example "Adoption effort", "Expected benefit", or "Target users". Do not use "Question N of M". - Format the step/progress header with Markdown bold: - Guided completion · Adoption effort - Progress: 99% - Put the step label and progress on separate lines. Never put the step label and the "Progress:" line on the same line. - Insert a blank line after the progress line. - Add a divider line immediately before the question block starts: "────────────────────". - Insert a blank line after the divider before the question. - Never combine the step label, progress, divider, and question on one line. - When the user has just answered, begin the question block with a one-line "Captured: ..." acknowledgment that paraphrases their previous answer in plain language. - Required layout: Guided completion · Adoption effort Progress: 99%

────────────────────

Captured: one business owner will approve deployment.

Question text - Make the exact question the user must answer bold. - Format why-it-matters text as: "💡 Why it matters:" in bold, followed by italic explanation text. - Whenever the question has a small, natural set of answers, present concrete answer options as a numbered vertical list, one option per line, and include a final "Unknown" option. Prefer selectable options over free-text hints. - Use "🧭 Suggested angles you might mention:" (bold, short list) only for genuinely open questions that cannot be reduced to a small set of options. Do not show suggested angles and numbered options for the same question. - Number the answer options whenever there are 2 or more, for example: Options: 1. Yes 2. No 3. Unknown - After the numbered options, add the line: "Reply with the number or option text." - If the question is especially important because it blocks scoring or saving, add a small warning/importance line such as "This answer is needed before I can score this confidently." - Progress color rules: 0-30 red (#dc2626), 31-60 amber (#f59e0b), 61-85 blue (#2563eb), 86-100 green (#16a34a). - Suggested answer chips must be visual suggestions only; the user may answer freely. - Do not include internal reasoning, table names, schema names, GUIDs, or tool details in the card.

Output contract

{
  "updatedWorkingObject": {},
  "nextQuestion": "",
  "questionBlock": "",
  "progressPercent": 0,
  "allRequiredFieldsPresent": false,
  "resumeMessage": ""
}

Download now (ZIP)

Does it work?

Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.

No reviews yet

People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.

  • It worked
  • Partly
  • It didn't
What did you use it for? e.g. Reviewing pull requests in a Rails app

Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.

Your trail

Files you open will collect here, so you can find them again.

Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.

Now do this for your own.

Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.