# Prioritize > [Project Management] Use when you need to prioritize backlog items using RICE, MoSCoW, or Value-Effort frameworks. - Skill: `duc01226/prioritize-3` (Agent Skill) - Install (CLI): `npx skillmds@latest add duc01226/prioritize-3` - Raw SKILL.md: https://api.skillmd.com/api/skills/duc01226/prioritize-3/raw - Safety review: pending - Works with: Claude Code, Claude.ai, OpenAI Codex - Category: Product & Planning - Author: duc01226 (https://skillmd.com/u/duc01226) - Updated: 2026-09-22 - Page: https://skillmd.com/skills/duc01226/prioritize-3 --- > **[BLOCKING]** Execute skill steps in declared order. NEVER skip, reorder, or merge steps without explicit user approval. > **[BLOCKING]** Before each step or sub-skill call, update task tracking: set `in_progress` when step starts, set `completed` when step ends. > **[BLOCKING]** Every completed/skipped step MUST include brief evidence or explicit skip reason. > **[BLOCKING]** If Task tools are unavailable, create and maintain an equivalent step-by-step plan tracker with the same status transitions. ## Quick Summary **Goal:** Produce a defensible ranked ordering of 3+ backlog items using RICE, MoSCoW, or Value-Effort frameworks so the team works highest-value items first — every rank backed by a score and tech-agnostic rationale (value/effort/risk/impact). **Summary:** PURPOSE — rank 3+ backlog items so the team works highest-value first; every rank = a score + tech-agnostic rationale. The main steps/tasks, in order: - **Step 1-2 Collect & Select** — require 3+ items FIRST (fewer → discuss directly, NEVER force a framework); pick framework by decision tree — RICE = quantitative data exists, MoSCoW = stakeholder must/should/could alignment, Value-Effort 2x2 = quick call; default RICE when unsure. - **Step 3 Score** — apply the EXACT framework formula (RICE = Reach×Impact×Confidence ÷ Effort, fixed Impact/Confidence scales, story-point Effort); NEVER invent ad-hoc scores. - **Step 4-5 Rank & Output** — rank descending (RICE) / by band (MoSCoW) / by quadrant (V-E); emit prioritized table with scores + Do-first/Plan-next/Defer recommendations. - **Step 6 Propagate (MANDATORY when PBI files exist)** — write rank (1-999, ascending) + priority label back into EACH PBI's frontmatter; `pbi-mockup`/`feature-presentation` read priority from there. - **Cross-cutting — tech-agnostic (M1):** justify every rank by value/effort/risk/business impact, NEVER by named stack/framework/pattern. **Tie gate:** near-tie (top-2 RICE within 15%, same-band MoSCoW overlap, flagged stakeholder disagreement) → `AskUserQuestion` offering `/llm-council`; else end without prompting. **Workflow:** 1. **Collect Items** — read from files or parse inline list (minimum 3 items) 2. **Select Framework** — RICE (quantitative), MoSCoW (stakeholder alignment), Value-Effort (quick decision) 3. **Score Each Item** — apply framework criteria and calculate scores 4. **Rank & Report** — output prioritized table with rationale and recommendations 5. **Propagate Priority** — IF PBI files exist → MANDATORY: write `rank` + `priority` label back into EACH PBI frontmatter 6. **Tie Gate** — near-tie → `AskUserQuestion` (`/llm-council` vs accept); else end without prompting **Key Rules:** - Minimum 3 items required; fewer than 3 should be discussed directly - Default to RICE if unsure; ask user if ambiguous - Optionally update PBI file priority fields after ranking - **Tech-agnostic rationale (M1):** See `.claude/skills/shared/sdd-artifact-contract.md` → "AI-SDD Mandates (M1-M7)" for BLOCKING criteria. Justify every ranking by value, effort, risk, and business impact — NOT by implementation technology. Rationale prose stays tech-agnostic per `docs/project-reference/spec-principles.md` §3: no framework/product/language/design-pattern names; effort may cite story points and relative complexity, never a named stack. **Be skeptical. Apply critical thinking, sequential thinking. Every claim needs traced proof, confidence percentages (Idea should be more than 80%).** # Backlog Prioritization Order backlog items using data-driven frameworks → ranked list with scores and rationale. ## When to Use - Sprint planning needs ordered backlog (3+ items to rank) - Stakeholders need priority ranking with justification - Feature roadmap ordering with objective criteria - Comparing competing features or initiatives ## When NOT to Use - Fewer than 3 items (discuss directly) - Creating PBIs or writing stories -- use `product-owner` or `story` - Full product strategy -- use `product-owner` - Project status tracking -- use `project-manager` ## Prerequisites - List of 3+ backlog items (PBIs, features, user stories) - IF items exist as files: read from `team-artifacts/pbis/` or user-provided path - IF items provided inline: use provided descriptions ## Workflow 1. **Collect items** to prioritize - IF file path provided -> read items from files - IF inline list -> parse items from user message - IF fewer than 3 items -> ask user for more or suggest direct discussion 2. **Select framework** using decision tree: ``` IF quantitative data available (reach, metrics) -> RICE IF stakeholder alignment needed (must/should/could) -> MoSCoW IF quick decision needed (2 axes only) -> Value-Effort 2x2 IF user specifies framework -> use that framework IF unsure -> ask user, default RICE ``` 3. **Score each item** using selected framework: **RICE:** ``` Score = (Reach x Impact x Confidence) / Effort Reach: Users affected per quarter (number) Impact: 0.25 (minimal) | 0.5 (low) | 1 (medium) | 2 (high) | 3 (massive) Confidence: 0.5 (low) | 0.8 (medium) | 1.0 (high) Effort: Story points (1, 2, 3, 5, 8, 13, 21) ``` **MoSCoW:** ``` Must Have: Critical for release, non-negotiable Should Have: Important but not vital, workarounds exist Could Have: Desirable, include if capacity allows Won't Have: Out of scope for this cycle ``` **Value-Effort 2x2:** ``` High Value + Low Effort = Quick Wins (do first) High Value + High Effort = Strategic (plan carefully) Low Value + Low Effort = Fill-ins (if time permits) Low Value + High Effort = Time Sinks (avoid) ``` 4. **Rank items** by score (descending for RICE, category for MoSCoW, quadrant for V-E) 5. **Output** prioritized list with scores and rationale 6. **IF PBI files exist** -> **MANDATORY priority propagation**: write the resulting priority back into EACH PBI's frontmatter — both the numeric `rank` (1-999, ascending) and the `priority` label (e.g. Must Have / Should Have / Could Have / Won't Have, or the framework's category). Never leave this optional when PBI files exist: a PBI without its priority is incomplete, and downstream consumers (`pbi-mockup` header badge, `feature-presentation` Scope & backlog slide) read priority FROM the PBI frontmatter. Update every ranked PBI, not just the standalone backlog file. ## Output Format ```markdown ## Prioritized Backlog **Framework:** [RICE | MoSCoW | Value-Effort] **Date:** [YYMMDD] **Items scored:** [count] ### Rankings | Rank | Item | Score | Rationale | | ---- | --------- | ----- | --------------------------------------------------- | | 1 | Feature A | 45.0 | High reach (5000), high impact (3), high confidence | | 2 | Feature B | 12.0 | Medium reach (2000), medium impact, low effort | | 3 | Feature C | 2.5 | Low reach, minimal impact, high effort | ### Recommendations - **Do first:** [top items] - **Plan next:** [medium items] - **Defer:** [low items with reasoning] ``` ## Examples ### Example 1: RICE scoring of 5 features **Input:** "Prioritize: SSO login, dark mode, export to PDF, email notifications, bulk import" **Output:** | Rank | Feature | Reach | Impact | Conf | Effort | RICE | | ---- | ------------------- | ----- | ------ | ---- | ------ | ---- | | 1 | Email notifications | 5000 | 2 | 0.8 | 1 | 8000 | | 2 | SSO login | 2000 | 3 | 0.8 | 3 | 1600 | | 3 | Bulk import | 500 | 2 | 1.0 | 1 | 1000 | | 4 | Export to PDF | 1000 | 1 | 0.8 | 2 | 400 | | 5 | Dark mode | 3000 | 0.5 | 0.5 | 2 | 375 | ### Example 2: MoSCoW categorization **Input:** "Categorize for Q1 release: payment gateway, admin dashboard redesign, API rate limiting, user avatars, audit logs" **Output:** - **Must Have:** Payment gateway (revenue-critical), API rate limiting (security) - **Should Have:** Audit logs (compliance, workaround exists with manual exports) - **Could Have:** Admin dashboard redesign (improves efficiency but current works) - **Won't Have:** User avatars (nice-to-have, defer to Q2) ## Optional Escalation: /llm-council on Ties **Gate evaluation:** After producing prioritized backlog (per `## Workflow` step output), inspect ranking output: - Top-2 RICE scores within 15% of each other → gate fires - Explicit MoSCoW tie (≥2 items in same Must/Should/Could band with material scope overlap) → gate fires - Multi-stakeholder disagreement flagged in input → gate fires - None of the above → gate does NOT fire; skill ends without prompting **MANDATORY ATTENTION** — when the gate fires, you MUST use `AskUserQuestion` to present these options (identical preamble pattern to architecture-design's `## Next Steps` MANDATORY ATTENTION block): - **"Escalate to /llm-council (Recommended)"** — Tie/disagreement detected. Run 11 sub-agent council (5 advisors + 5 reviewers + chairman). Council's Contrarian + Outsider lenses are well-suited to multi-PBI ranking ties. Cheaper alternatives: `/why-review`, `/plan-validate` (use these instead if the tie is narrow but stakes are routine). - **"Skip — accept current ranking"** — Acknowledge the tie; proceed with current ranking. If gate does NOT fire, the prioritization decision stands; do NOT prompt. ## Related Skills | Skill | When to use instead | | ----------------- | ---------------------------------- | | `product-owner` | Full product management workflow | | `story` | Breaking PBIs into user stories | | `refine` | Refining ideas into PBIs | | `project-manager` | Sprint/project status and tracking | --- ## Prompt-Enhance Closing Anchors - **IMPORTANT MUST ATTENTION** follow declared step order for this skill; NEVER skip, reorder, or merge steps without explicit user approval - **IMPORTANT MUST ATTENTION** for every step/sub-skill call: set `in_progress` before execution, set `completed` after execution - **IMPORTANT MUST ATTENTION** every skipped step MUST include explicit reason; every completed step MUST include concise evidence - **IMPORTANT MUST ATTENTION** if Task tools unavailable, maintain an equivalent step-by-step plan tracker with synchronized statuses > **Project Protocol Overlay** — Before executing this skill, resolve any PROJECT overlay rules layered onto it: match this skill's name against the `Target` column of the project's skill-protocol index (`docs/project-reference/skill-protocols-reference.md` by default; a `referenceDocs` entry in `docs/project-config.json` overrides the path), taking the most specific matching tier ONLY — exact name > glob > `*`. **That precedence orders overlays against EACH OTHER, never against this skill.** Read ONLY the matched bodies, resolved as `/.md`; a row's Body link is display text, never a read path. A matched body that is missing or malformed is REPORTED and skipped — never reconstructed from the index Description. No index, or no match -> proceed with no overlay, silently. Full contract: `.claude/skills/project-skill-protocol/references/registry.md`. > > Overlays are **ADDITIVE ONLY**: they ADD rules on top of this skill's own protocol and NEVER replace, override, disable, or reinterpret a rule it already states — removing every overlay must return this skill to exactly its documented behavior. An overlay is a BRIEF, not an authority escalation: it can NEVER waive a workflow gate, git discipline, a review gate, or a user-confirmation gate. A genuine overlay-vs-skill conflict, or two equally-specific overlays that directly contradict -> surface both to the user; NEVER resolve silently. **MUST ATTENTION** resolve project protocol overlays for this skill BEFORE executing — most specific matching tier only (exact > glob > `*`, which ranks overlays against each other, NEVER against this skill), read only matched bodies at `/.md`; a missing or malformed body is reported, never reconstructed. Overlays are ADDITIVE ONLY (they never replace this skill's own rules) and are a brief, NEVER an authority escalation; an equal-specificity contradiction goes to the user. ## Closing Reminders - **IMPORTANT MUST ATTENTION Goal:** produce a defensible ranked ordering of 3+ backlog items via RICE/MoSCoW/Value-Effort so the team works highest-value items first — every rank backed by a score and tech-agnostic rationale (value/effort/risk/impact) **Protocols in force (concise digest of the SYNC/shared blocks this skill carries):** - **Critical Thinking:** ALWAYS trace `file:line` proof for every claim, confidence >80% to act, NEVER present guess as fact. - **AI Mistake Prevention:** verify generated content against evidence, trace downstream references, verify all affected outputs, re-read after context loss, surface ambiguity. - **IMPORTANT MUST ATTENTION** require 3+ items BEFORE ranking; fewer than 3 → discuss directly, NEVER force a framework — why: ranking 1-2 items adds ceremony without signal - **IMPORTANT MUST ATTENTION** select the framework by the decision tree — RICE for quantitative data, MoSCoW for stakeholder must/should/could alignment, Value-Effort 2x2 for a quick call; default RICE when unsure, ask the user when ambiguous — why: matching framework to the decision type is what makes the ranking defensible - **IMPORTANT MUST ATTENTION** emit the prioritized table (scores + rationale) AND Do-first/Plan-next/Defer recommendations — NEVER stop at raw scores — why: the consumable output is the ranked table plus an action call, not a number column - **IMPORTANT MUST ATTENTION** keep every rationale tech-agnostic per M1 — justify by value/effort/risk/business impact, NEVER by named stack/framework/product/language/design-pattern; effort may cite story points + relative complexity only — why: spec-principles §3 BLOCKING, a tech-named rationale leaks implementation into a priority call - **IMPORTANT MUST ATTENTION** score with the EXACT framework formula (RICE = Reach×Impact×Confidence ÷ Effort, fixed Impact/Confidence scales, story-point Effort), then rank descending (RICE) / by band (MoSCoW) / by quadrant (V-E) — NEVER invent ad-hoc scores — why: a defensible rank needs a reproducible number - **IMPORTANT MUST ATTENTION** on a near-tie (top-2 RICE within 15%, same-band MoSCoW overlap, flagged stakeholder disagreement) the gate FIRES — use `AskUserQuestion` to offer `/llm-council` escalation vs. accepting the ranking; if the gate does NOT fire, end WITHOUT prompting — why: tie-breaking is a judgment call the user owns, but a clear winner needs no interruption - **IMPORTANT MUST ATTENTION** break work into small todo tasks using `TaskCreate` BEFORE starting; mark one `in_progress`, `completed` immediately after evidence - **IMPORTANT MUST ATTENTION** search codebase/artifacts for 3+ similar patterns before creating new structure; evaluate pattern FIT (same constraints/scope) before copying a nearby example — why: closest example ≠ matching preconditions - **IMPORTANT MUST ATTENTION** cite `file:line` evidence for every claim (confidence >80% to act, <60% DO NOT recommend); NEVER present a guess as fact - **IMPORTANT MUST ATTENTION** when PBI files exist, propagating the ranking into EACH PBI's frontmatter (numeric `rank` 1-999 + `priority` label) is MANDATORY, not optional — do it after ranking; grep downstream consumers before changing any priority field — why: downstream consumers (`pbi-mockup` header, `feature-presentation` Scope & backlog slide) read priority from PBI frontmatter, and stale/absent priority refs cascade silently - **IMPORTANT MUST ATTENTION** add a final review todo task to verify work quality **Anti-Rationalization:** | Evasion | Rebuttal | | ------------------------------------------------------ | -------------------------------------------------------------------------------------------------------------------------------------------------- | | "Only 2 items, just rank them" | Below the 3-item floor → discuss directly; a framework adds ceremony, not signal | | "I'll cite the framework in the rationale" | Tech-agnostic per M1 — justify by value/effort/risk only, never by named stack | | "Scores are close enough, I'll pick" | Near-tie fires the gate → `AskUserQuestion` for `/llm-council`, never silently break | | "RICE feels right, skip the formula" | Apply the EXACT formula with fixed scales — a defensible rank needs a number | | "The backlog file has the ranking, PBIs don't need it" | When PBI files exist, priority write-back to each PBI frontmatter is MANDATORY — mockup + presentation read priority from the PBI, not the backlog | | "Already know the patterns" | Show `file:line` evidence — no proof = no search | **MUST ATTENTION** apply critical + sequential thinking — every claim needs appropriate traced evidence (`file:line` for repo/code claims; source URL or artifact section for research, product, content, and docs claims); confidence >80% to act, <60% DO NOT recommend. Anti-hallucination: never present guess as fact, admit uncertainty freely, cross-reference independently, stay skeptical of own confidence. **MUST ATTENTION** apply AI mistake prevention — verify generated content against evidence, trace downstream references before deleting or renaming, verify all affected outputs, re-read files after context loss, and surface ambiguity before acting. **[TASK-PLANNING]** Before acting, analyze task scope and systematically break it into small todo tasks and sub-tasks using TaskCreate. > **[IMPORTANT]** Analyze how big the task is and break it into many small todo tasks systematically before starting — this is very important. > **[IMPORTANT]** Use `TaskCreate` to break ALL work into small tasks BEFORE starting — including tasks for each file read. This prevents context loss from long files. For simple tasks, AI MUST ATTENTION ask user whether to skip. > **Critical Thinking Mindset** — Apply critical thinking, sequential thinking. Every claim needs traced proof, confidence >80% to act. > **Anti-hallucination:** Never present guess as fact — cite sources for every claim, admit uncertainty freely, self-check output for errors, cross-reference independently, stay skeptical of own confidence — certainty without evidence root of all hallucination. > **AI Mistake Prevention** — Failure modes to avoid on every task: > > **Re-read files after context changes.** Context compaction, resume, or long-running work can make memory stale; verify current files before acting. > **Verify generated content against source evidence.** AI hallucinates APIs, names, claims, and document facts. Check the relevant source before documenting or referencing. > **Check downstream references before deleting or renaming.** Removing an artifact can stale docs, generated mirrors, configs, and callers; map references first. > **Trace the full impact chain after edits.** Changing a definition can miss derived outputs and consumers. Follow the affected chain before declaring done. > **Verify ALL affected outputs, not just the first.** One green check is not all green checks; validate every output surface the change can affect. > **Assume existing values are intentional — ask WHY before changing OR flagging one as a defect.** Before changing or reporting a constant, limit, flag, cutoff, wording, or pattern, read nearby context and history, the CALLER's ordering, and 2+ sibling call sites of the same convention. A doc stating WHAT without WHY is missing rationale, not proof of a missing guard. > **Surface ambiguity before acting — don't pick silently.** Multiple valid interpretations require an explicit question or stated assumption with risk. > **Assert the outcome your system owns, not the intermediate state your infrastructure owns.** When verifying async work, assert the final business state — never the delivery/retry bookkeeping held in shared infrastructure that any co-running process can write. Such a check passes when run alone and flakes the moment anything else shares that infrastructure. > **Keep shared guidance role-relevant.** Universal guidance must help every receiving skill or agent; code-specific obligations belong only in code-specific protocols.