daddia Design Decision Rights
When to use
Assign RACI/RAPID roles with exactly one Accountable per decision — enforce single-point accountability and flag Consulted bloat.
What this skill does not do
- Does not diagnose structure fit — route to
/operating-model:diagnose-structure-fit. - Does not resolve political ownership — forces a choice; strategist decides.
- Does not implement governance systems — produces artifact for adoption.
Preconditions
| Input | If missing |
|---|---|
| Decisions or processes in scope | Ask user to list |
| Practice profile (RACI vs RAPID convention) | Default RACI; flag [PROVISIONAL] |
| Named roles/parties for assignment | Ask; halt if decisions unnamed |
Provisional mode
Multiple Accountable proposed: flag diffused accountability; do not finalize until single owner chosen or explicitly [review].
Trust spine
- Confidence bands (
structured-aggregation):- High: Every decision has exactly one Accountable; bloat flags applied.
- Medium: Some ownership gaps with recommended owners.
- Low: Zero Accountable on key decisions — halt output until resolved.
- Failure modes:
- Strategic advice vs. support: Assignments are draft for exec confirmation.
- Client confidentiality: RACI may expose sensitive authority lines — CONFIDENTIAL header.
- Accountability gap: "Shared ownership" rejected; forces single Accountable.
- Analytical Rigor: MECE decision coverage in scope.
- Incentive Gaming: Consulted bloat flagged as latency/gaming risk.
- Escalation triggers: Unowned decisions found — name specifically, do not defer.
Workflow
- Read practice profile for decision-rights gaps and house convention.
- Assign roles per decision — R/A/C/I or RAPID equivalent.
- Enforce exactly one Accountable — stop if zero; flag if multiple.
- Check Consulted bloat — long lists that slow decisions materially.
- Find unowned decisions — no claimed owner during exercise.
- Completeness check before output.
Output format
CONFIDENCE: [defensible recommendation | structured first pass]
DECISION: [name]
Accountable: [single name/role — flag if zero or multiple initially proposed]
Responsible: [...]
Consulted: [...] — Bloat flag: [if applicable]
Informed: [...]
[repeat per decision]
UNOWNED DECISIONS FOUND: [list or none]
Worked example
Input: "Product pricing changes" — Accountable: Product VP and Finance VP.
Expected output (excerpt):
DECISION: Product pricing changes
Accountable: FLAG — multiple proposed (Product VP, Finance VP); force choice [review]
UNOWNED DECISIONS FOUND: none
Quality checks before delivering
- Exactly one Accountable per decision (or explicit flag)
- Consulted bloat assessed
- Unowned decisions listed
- House convention (RACI/RAPID) applied
- Output does not read as final governance policy
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/operating-model/CLAUDE.md (org-wide facts go to org-profile.md), ask for confirmation, and write only on yes. Only /operating-model:practice-setup auto-applies a full profile write.
Outputs
Follows plugin CLAUDE.md § Outputs. Next: exec workshop to resolve flags, socialize artifact, or stress-test-matrix-reporting.