| name | narrative-builder |
| description | This skill should be used when the user asks to "build a narrative from these notes," "what's my storyline here," "structure this into a pyramid," or provides raw findings, interview notes, or analysis output and wants it turned into a governing thought with supporting argument structure.
|
| allowed-tools | Read, Grep, Glob |
| metadata | {"version":"0.3.0","owner":"consulting practice","review_cadence":"quarterly","work_shape":"narrative-synthesis","permission_tier":"advisory","output_class":"draft-for-review","sourcing_policy":"volatile-facts-must-be-sourced"} |
Narrative Builder
When to use
Turn raw input into Minto pyramid: governing thought, 3–5 MECE key line points, support bullets. See ../../references/minto-pyramid.md.
What this skill does not do
- Does not produce prose memo/deck — skeleton for
deck-outline / exec-memo.
- Does not paper over gaps — flags unsupported governing thoughts.
- Does not guess reader question — ask if unclear.
Preconditions
| Input | If missing |
|---|
| Raw findings/notes | Ask user to provide |
| Reader's real question | Ask before structuring |
| Practice profile | Default inductive; flag practice-setup |
Provisional mode
Thin input: governing thought flagged [review]; Known gaps section required.
Trust spine
- Analytical Rigor (mandatory): Outline not prose; MECE gaps/overlaps named; grouping logic stated.
- Per
trust-conventions.md.
Workflow
- Read practice profile.
- State reader's question explicitly.
- Draft governing thought (answer, not topic).
- 3–5 MECE key line points; state deductive vs inductive.
- Support sub-bullets with so-what test.
- Flag structural problems in source — don't silently fix.
Output format
GOVERNING THOUGHT: [...]
1. [Key line point 1]
- [Support]
2. [...]
Grouping logic: [deductive | inductive] — [rationale]
Known gaps / flags: [...]
Worked example
Input: Interview notes on pricing. Question: should we revert mid-tier increase?
Excerpt: Governing thought answers revert yes; three MECE key lines; gap flagged: no competitor primary research.
Quality checks before delivering
Propose profile update
When a stable convention surfaces during this run (thresholds, naming, tone, output format, or recurring corrections), propose a profile update: show the exact diff against ~/.claude/plugins/config/claude-for-strategy/consulting/CLAUDE.md (org-wide facts go to org-profile.md), ask for confirmation, and write only on yes. Only /consulting:practice-setup auto-applies a full profile write.
Outputs
Follows plugin CLAUDE.md § Outputs. Next: /consulting:deck-outline or /consulting:exec-memo.
