--- name: weekly-pipeline-digest description: Generates a weekly pipeline digest from HubSpot deals — what moved, what stalled, and the top deals to push this week — formatted for email or a Monday standup. Use when the user asks for a "pipeline digest", "Monday pipeline brief", "what changed in pipeline this week", or sets up a recurring weekly pipeline summary. license: MIT compatibility: Requires a connected HubSpot account. Read-only. metadata: author: Builder.org version: 1.0.0 persona: Operator Owen, Strategic Steve rave: R4 A4 V4 E4 category: hubspot --- # Weekly Pipeline Digest ## What it does A read-only Monday-morning digest of the open pipeline: totals, movement, stalls, and a short "push list" — so the user is caught up before the team standup. ## When to use - A scheduled weekly run, or on request: "give me the pipeline digest", "what moved in pipeline this week". ## Inputs None required. Optionally a pipeline name, owner, or segment to scope it. ## Workflow 1. **Snapshot now.** Search open deals; capture totals (count, value), per-stage counts/value, and each deal's stage + days-in-stage. 2. **Compare to last week.** Retrieve the prior weekly snapshot if one exists; compute deltas (pipeline value, new deals, deals advanced, deals slipped). On the first run, state there's no prior baseline and skip deltas. 3. **Flag stalls.** Mark deals stuck 14+ days in stage with an SLA cue. 4. **Build the push list.** Pick the top 3 deals to act on this week (closest to close, highest value, or most at-risk) with a one-line reason each. 5. **Write the digest** using the template. Keep it skimmable. 6. **Persist the snapshot** for next week's comparison (store the compact totals + per-stage map only). ## Output ``` ## Pipeline digest — week of [date] **Open pipeline:** $[X] across [N] deals ([+/- vs last week]) **New this week:** [N] · **Advanced:** [N] · **Slipped:** [N] ### Stalled (14+ days in stage) - [Deal] — [stage], [days] days idle ### Top 3 to push this week 1. [Deal] — [why] 2. … 3. … ``` ## Notes - Read-only by design: it reports, it doesn't change deals. - Reliable across the 80% case because it works on aggregate + stage data, not fragile per-field parsing.