emmy-requirements-context

name	emmy-requirements-context
description	Load active Emmy planning requirements from the Goals Space requirements layer. Use when Emmy planning, ticket proposals, release readiness, architecture decisions, or goal/outcome alignment need requirement constraints, blockers, due dates, owners, or evidence links.

Emmy Requirements Context

Use this skill to load requirement constraints that shape Emmy planning work. Requirements are part of the Goals Space intent layer: they say what must remain true, what blocks work, and what evidence is expected before a goal, outcome, ticket, release, or architecture decision can move forward.

This v1 skill is read-only. It reads the Goals Space Requirements Registry page when present and returns an empty result with a warning if the registry is missing or unreadable.

Source Of Truth

• Goals Space root: 1398768351
• Requirements Registry: 1399165621
• Requirements Registry title: Requirements Registry
• Confluence space: SFIV
• MCP server: cms-atlassian-confluence
• Required user environment variable: CONFLUENCE_PERSONAL_TOKEN
• Shared plugin MCP config may expose Jira tools. This skill uses Confluence only.

This skill complements emmy-planning-context. Do not write to the Goals Space, do not create Jira tickets, and do not dereference Knowledge Store evidence links. Return Knowledge Store links as links only.

Inputs

Accept these caller options when supplied:

Name	Type	Default	Meaning
focus_area	string | null	null	Optional topic hint used to narrow requirement pages.
goal_ids	string[]	[]	Goal IDs whose related requirements should be loaded.
outcome_ids	string[]	[]	Outcome IDs whose related requirements should be loaded.
principle_ids	string[]	[]	Principle IDs whose related requirements should be loaded.
include_inactive	boolean	false	Whether to include Deferred, Retired, or Superseded requirements.

Read Sequence

1. Read Goals Space root 1398768351.
   • Verify the title is Emmy Goals Space.
   • Stop with a hard error if the root is missing, unreadable, or uninitialized.
2. Read Requirements Registry 1399165621 and verify the title is Requirements Registry.
   • If direct lookup fails, read root children and search for exact title Requirements Registry.
3. If no registry exists or no readable registry can be found, return:
   • requirements: []
   • warning code REQUIREMENTS_REGISTRY_MISSING
   • metadata showing the Goals Space root and children were read.
4. Read the Requirements Registry page.
   • Parse the active requirements table when present.
   • Ignore placeholder rows such as -, --, ---, None yet, or em dash equivalents.
   • Include only rows with status Active unless include_inactive is true.
5. For each matching requirement row, read the linked child requirement page in full.
   • Apply focus_area, goal_ids, outcome_ids, and principle_ids case-insensitively against the requirement name, domain, planning impact, blocking rule, related IDs, open questions, and link text.
   • If no requirements match the supplied filters, return an empty list plus a warning code NO_MATCHING_REQUIREMENTS.
6. Assemble and return RequirementsContext.

RequirementsContext Shape

Return structured data rather than a prose summary:

{
  "requirements": [
    {
      "id": "REQ-ATO-001",
      "name": "Maintain valid ATO",
      "domain": "security-authorization",
      "status": "Active",
      "owner": "ISSO / System Owner / Emmy Team",
      "due_or_review_date": "2026-09-21",
      "authority": "Emmy ATO Tracker",
      "applicability": "Production operation and release planning",
      "planning_impact": "Tickets that create new external interfaces need approval before implementation.",
      "blocking_rule": "No new non-CMS interfaces without prior approval.",
      "related_goals": ["GOAL-001"],
      "related_outcomes": ["GOAL-001-O2"],
      "related_principles": ["PRIN-004", "PRIN-005"],
      "evidence_links": [
        {
          "label": "Emmy ATO Tracker",
          "url": "https://confluenceent.cms.gov/pages/viewpage.action?pageId=1112978748"
        }
      ],
      "open_questions": [],
      "source_page_id": "1390000000",
      "source_url": "https://confluenceent.cms.gov/pages/viewpage.action?pageId=1390000000"
    }
  ],
  "warnings": [
    {
      "code": "REQUIREMENTS_REGISTRY_MISSING",
      "page_id": "1398768351",
      "message": "No Requirements Registry child page exists under the Goals Space root."
    }
  ],
  "metadata": {
    "loaded_at": "2026-06-16T12:00:00Z",
    "pages_read": ["1398768351"],
    "page_versions": {
      "1398768351": 1
    },
    "focus_area": null,
    "requirements_registry_page_id": "1399165621",
    "goals_space_url": "https://confluenceent.cms.gov/spaces/SFIV/pages/1398768351"
  }
}

Always populate metadata.pages_read with every page ID consulted.

Requirement Page Expectations

When requirement child pages exist, extract these sections when present:

• Status, Owner, Domain, Authority, Due or review date.
• Requirement statement.
• Applicability trigger.
• Planning impact.
• Blocking rule.
• Evidence expectation.
• Related goals, outcomes, and principles.
• Evidence links.
• Open questions.
• Change log.

If a section is missing, preserve the requirement and add a targeted warning instead of inventing a value.

Error And Warning Rules

• Missing or uninitialized Goals Space root: stop with a hard error naming page 1398768351.
• Missing Requirements Registry: return warning REQUIREMENTS_REGISTRY_MISSING and continue with empty requirements.
• Empty registry table: return warning REQUIREMENTS_REGISTRY_EMPTY.
• No matching filtered requirements: return warning NO_MATCHING_REQUIREMENTS.
• Unreadable requirement child page: add warning UNREADABLE_REQUIREMENT_PAGE, skip that requirement, and continue.
• Missing expected fields on a requirement page: add warning REQUIREMENT_FIELD_MISSING.

Constraints

• Read only. Do not create, update, retire, or reorder requirements.
• Do not create, update, or transition Jira tickets.
• Do not append to the Planning Log.
• Do not dereference Knowledge Store links or call Knowledge Store skills.
• Do not treat the absence of a registry as proof that no requirements apply. Report the missing registry warning.
• Cite page titles and URLs or page IDs when requirement context materially shapes the answer.

MCP Tool Use

Use Confluence MCP tools such as confluence_get_page, confluence_get_page_children, confluence_search, confluence_get_page_history, and confluence_get_page_diff for reads.

Prefer convert_to_markdown: true for normal extraction. Use storage format only when markdown conversion hides table or link details needed to parse the context accurately.