/lynqu-pipeline-report

name: lynqu-pipeline-report
description: Weekly Lynqu pipeline and sales report. What moved, what's stalling, campaign and event ROI, card engagement, who needs attention. Read-only. Requires the Lynqu MCP server connected.

Lynqu Pipeline Report

You produce the briefing a sales leader reads on Monday morning: what moved, what didn't, what's at risk, and the three things to do about it. **This skill is read-only.** It never writes, never sends, never moves a stage. That constraint is what makes it safe to run on anything, and it's why it's the natural last step after every other skill in the suite.

A report that lists numbers is a dashboard, and the user already has one. A report that says *"the pipeline looks healthy but 60% of it is one deal that hasn't moved in five weeks"* is worth the read.

Invocation

/lynqu report [week | month | quarter | board | rep | campaign | event]

Defaults to the last 7 days across the whole org, scoped to what the caller's role can see.

Step 1: Gather

Pull broad first, then drill only where something looks wrong.

| Source | What it gives you | |--------|-------------------| | `get-org-summary` | The shape: counts by stage, activity, campaigns | | `get-dashboard-summary` | Headline numbers, pre-aggregated | | `list-leads` (by stage, updated window) | Movement and staleness. The raw material | | `get-forecast` | Weighted pipeline by forecast category (manager+) | | `list-campaigns` + `get-campaign` | Spend against return | | `get-event-portfolio` | Cross-event roll-up: leads, pipeline, won, cost | | `list-events` | What's live and what's closing | | `get-team-performance` / `get-employee-performance` | Per-rep (add-on gated) | | `get-card-stats` / `list-org-cards` | Card engagement: the top of the funnel | | `list-dashboards` + `get-dashboard-data` | The org's own saved widgets (analytics add-on) | | `list-lead-duplicates` | Data quality that's distorting every number above | | `get-followup-performance` | Which templates, sequences, steps and reps get replies, meetings and wins | | `list-outbox` | What actually went out: bounces, server cancellations, and the replies that came back | | `list-handoffs` | Work AI employees handed back that nobody has picked up (only if the org has them) | | `list-segments` + `get-segment-performance` | Which market, source or cohort converts: win rate, hours to first response, reply rate, meetings, per segment, rows adding up (analytics add-on) |

Skip cleanly on `ADDON_REQUIRED`. Analytics dashboards and team performance are add-ons; the core report needs neither. Mention the omission once, at the end, without turning it into an upsell.

Step 2: Analyse differences, not totals

A total is a fact. A difference is information.

1. **Movement.** What entered, advanced, was won, was lost. Compare with the previous equivalent period. "12 new leads" means nothing; "12 new, against a 4-week average of 31" is the headline. 2. **Concentration.** How much of the pipeline is one deal, one rep, one campaign? This is the risk nobody puts on a dashboard, and it's usually the real story. 3. **Stalls.** Leads past 2× the median time in stage, and where they pile up. A specific stage that always clogs is a process problem, not a rep problem. 4. **Velocity.** Time from creation to won, and whether it's changing. 5. **Conversion by stage.** Where deals actually die. If half of Evaluation goes to lost, the problem started earlier, in qualification. 6. **Source quality.** Win rate by capture source and campaign. Volume flatters the wrong channel constantly. 7. **Coverage.** Weighted pipeline against target. Say the ratio plainly. 8. **Follow-up engagement.** `get-followup-performance` by template, then by sequence step (`group: step` with the sequence's `trigger_id`) to find where a sequence goes quiet. Reply rate is the headline; a row with fewer than 20 delivered emails is unranked, so report its numbers without a verdict. When `replies_tracked` is false, replies are not being collected: say reply rates are unmeasured rather than reporting 0%, and lean on clicks and meetings booked. Open rate is empty when the org does not track opens. 10. **Segment quality.** `get-segment-performance` over the same window. Each lead sits in the first segment it matches, so the rows add up to the org; a segment is ranked by win rate only past 20 decided deals, and the `segment:former` row is email to a deleted segment, not a market. Money cells carry an `fx` block when currencies were converted or left out: say so. Skip cleanly on `ADDON_REQUIRED`. 9. **Hygiene.** Duplicates, orphaned leads, open leads with no next step. These are the caveats every other number in the report needs.

Step 3: Turn analysis into recommendations

Every finding needs a **so what**. Three to five recommendations, ranked by impact, each naming a specific action, a specific target, and who does it.

Bad: "Follow-up could be improved." Good: "9 hot leads from SaaStr have had no touch in 6 days. The 48-hour window is gone. Run `/lynqu followup` on that cohort today; expected recovery is 2–3 meetings based on the last event's rate."

Rank by revenue impact, not by how easy the fix is.

Step 4: Say what you can't see

Every report states its own limits. This is not a disclaimer, it's the part that makes the rest trustworthy:

• Which surfaces were unavailable (add-ons, role scope)
• Data quality problems that distort the numbers. Duplicates double-count value,

leads with no value set are invisible in the forecast

• Where the sample is too small for the percentage to mean anything. "3 of 4 = 75%

win rate" is not a win rate

• Whether replies are tracked at all (`replies_tracked`). An untracked reply

rate is not a zero

A leader who catches a report overstating its confidence once will discount every report after it.

Output format

# Pipeline Report · {period}

**{One-sent