← ClaudeAtlas

prioritizelisted

General backlog prioritization — pulls the backlog from Jira, scores each item with RICE or ICE, ranks them, and recommends what to build next. Scores are tied to the business's own metrics from business.json (north star, key metrics, revenue model), assumptions are stated explicitly, and unknowns are confirmed with you rather than invented. Optionally writes scores back to Jira and publishes a summary to Confluence (draft-before-publish). Use when asked to 'prioritize backlog', 'rank the backlog', 'what should I build next', 'rice scoring', 'ice scoring', or '/prioritize'. (Distinct from /audit-offers, which is an offer/discount ROI audit.)
mohitkhandelwal242/ai-pm-operator · ★ 0 · AI & Automation · score 65
Install: claude install-skill mohitkhandelwal242/ai-pm-operator
# /prioritize — Backlog Prioritization (RICE / ICE) Pull the backlog from Jira, score every candidate with a transparent prioritization framework, rank them, and tell the user what to build next — with each recommendation tied to the business's own metrics. **Iron Law: NEVER INVENT NUMBERS.** Every Reach / Impact / Confidence / Effort value is derived from a real signal (story points, labels, issue type, `business.json`) or confirmed by the user. State every assumption inline. A score with a fabricated input is worse than no score. --- ## Step 0: Bootstrap Read silently: 1. `.env` to load: - `JIRA_PROJECT_KEY` (referred to as `PROJECT_KEY` below) - `CONFLUENCE_SPACE_KEY`, `CONFLUENCE_PARENT_PAGE_ID`, `CONFLUENCE_CLOUD_ID` 2. `business.json` — required context. Pull out: - `metrics.north_star` — the single metric the ranking should serve. - `metrics.key_metrics[]` — the metrics Impact is measured against. - `company.revenue_model` — `subscription`, `transaction`, `ads`, or a mix; shapes how Impact is interpreted. - `company.target_users` — used for Reach sizing. 3. `team.json` — roster, to size Effort against real capacity and name owners in recommendations. 4. Parse `$ARGUMENTS`: - `--framework rice|ice` — scoring model. Default: `rice`. - `--top N` — how many items to recommend. Default: `5`. - `--epic KEY` — scope the backlog to one epic's children only. - `--no-publish` — console output only; never touch Jira or Confluence. Constants: -