ashokraj2011

sflow-epic-requirements

Formalize pinned Epic sources into requirements, traceability, and repository impact analysis, then publish one governed Requirements bundle.

ashokraj2011 0 Updated 2w ago
GitHub

Install

npx skillscat add ashokraj2011/singularityflow/sflow-epic-requirements

Install via the SkillsCat registry.

SKILL.md

Build the Epic Requirements bundle

Output contract: Use governed inputs and pinned clarification; publish/show configured artifacts. For suggested actions, pair Shell with the returned Copilot command; honor commandGuidance. If absent, say "Copilot: no verified equivalent"; never invent a slash command.

Boundary: no Story required; cwd=opened Git root or verified repositoryPath from singularity-flow workspace current --json; refuse if neither resolves; never search $HOME/parents.

  1. Run singularity-flow epic requirements status --json. Epic Intake accepts the Jira identity and pinned sources automatically; do not build or request a repository world model on main or the Epic branch.
  2. List pinned inputs with singularity-flow epic sources list --epic <EPIC-KEY>.
  3. Record every business note with singularity-flow epic sources note --text-file <FILE> and every accepted Copilot answer with singularity-flow epic sources answer --text-file <FILE>. Never leave accepted answers only in chat history.
  4. Run singularity-flow epic requirements prepare. Read the complete governed context that it prints: phase contract, governed agent, Jira snapshot, pinned sources, question ledger, and workspace repository registry. Repository grounding is added later, after each Story has its canonical branch.
  5. Use ask_user to ask one concise batch of material questions and wait. Even when the pinned evidence looks complete, ask the contributor to confirm your concise interpretation of outcome, boundaries, and acceptance criteria. If ask_user is unavailable, display the questions and stop before authoring or publication. Record every accepted answer with singularity-flow epic sources answer, rebuild context, and continue. Unanswered questions stay visible; never fabricate scope.
  6. Author all three outputs: requirements.md, traceability.yml, and impact-analysis.yml. Use stable REQ-nnn and AC-nnn identifiers and cite exact SRC-nnn evidence.
  7. Run singularity-flow epic requirements publish, show all three documents and hashes in full, and stop for review.
  8. After explicit approval, run singularity-flow epic requirements approve. Never approve automatically.