SuperAPI AI Directory
Partner中文AI newsFavorites
Skillsteam-coaching-rollup

team-coaching-rollup

Office & productivity

View repository

Updated 10-16 08:45

About this skill

Team Coaching Rollup Skill - v1.0.0

<run_context> Audience: Tim and Victor. This is the trainer and manager rollup over the whole team, not a rep self-serve tool. It resolves whoever runs it (for the optional Slack DM) but it always reports across all four reps.

  • Running user = the person in this Claude's personalization (Work Email); resolve from the roster. Intended runners are Tim (87486452) and Victor (153471050).

  • Team roster (HubSpot owner id + Nooks user id, scope each rep's pull by these):

    Rep Role Email HubSpot owner_id Nooks user id
    Tim Kipper Sr BDR tkipper@epiphan.com 87486452 UdhD9yeRfOXij75iD0Oy9F1b25P2
    Edgar Marroquin SDR emarroquin@epiphan.com 93367782 U2iWVwWG83OYhm3XR0mvGd9gWDV2
    Vasil Ivanov SDR vivanov@epiphan.com 93782443 1BUM3d2fZuhsbFttl5UOKEOPuMG3
    Nyasha Chigwedere SDR nchigwedere@epiphan.com 94135434 LGCdlTv6v8dSAiSsoJ7ETWQHVKl1
  • Manager: Victor Doubrovine (153471050). AEs (for the demos-booked read): Phil Sandler (190030668), Lex Evans (82625923).
  • Geography framing: USA and Canada team. Rest-of-World is Taroni's book (165812850); it is not part of this team rollup.
  • MCP namespace: default Epiphan Ai; fall back to Epiphan CRM only for tools missing on the subset. Nooks server 8c1d2738-2529-4248-b268-342859275d11; Nooks READ tools only (getMe, listCalls, listCallDispositions, getCallDisposition, listTasks, listUsers); every Nooks call carries USER_GOAL + a per-call REASON.
  • Read-only everywhere except one optional Slack DM to the running user. No HubSpot, CRM, or Nooks record is ever written. </run_context>

<data_sources> Pull per rep, scoped by that rep's ids. Never blend reps; scope every query.

  • HubSpot rep_activity via query_dataset (rep_activity), scoped by each rep's HubSpot owner id: dials and connects (raw volume) and demos booked in the period. This is the source of truth for volume and connects.
  • Nooks listCalls and listCallDispositions, scoped by each rep's Nooks user id: the disposition mix (connected, no answer, voicemail, callback, not interested, and so on, using the live disposition names) and callbacks owed. This is the source for the mix and callbacks, not for raw totals.
  • Reconcile like the morning brief: when HubSpot and Nooks disagree on counts, HubSpot raw wins for volume and connects; Nooks is used only for the disposition mix and callbacks. Note any large gap between the two as a data-hygiene flag rather than silently picking one.
  • Demos booked: from HubSpot for the period (meetings or demo-stage deals scoped to the rep's owner id), cross-checked against any AE handoffs. State the basis you used. </data_sources>

<coaching_flags> Flags are signals for a conversation, not verdicts. Raise the ones that apply and pair each with a concrete, humane focus, not a scold. Defaults are tunable in <config>.

  • Low connect rate: connect rate below the team band or well below the team median. Focus: timing (pair with nooks-call-timing) and list quality, not just "dial more".
  • Weak disposition hygiene: a high share of calls with no disposition or a vague catch-all, or a big HubSpot-vs-Nooks gap. Focus: clean dispositioning so the data is trustworthy (pair with nooks-call-wrap).
  • No demos: activity but zero demos booked in the period. Focus: opener and discovery quality (pair with sdr-ramp-cert and the call playbook), not volume.
  • Heavy no-answer: no-answer share well above the team median. Focus: call windows and multitouch on the no-connects (pair with no-connect-multitouch).
  • On pace / strong: call it out too. Coaching is not only about gaps; name what is working so it spreads. Each rep gets at most the two or three flags that matter most, with one suggested focus, so the read is usable rather than a wall of warnings. </coaching_flags>

This skill produces no prospect-facing copy and states no product specs, so GATE B (technical accuracy) and GATE C (brand voice) do not apply. If a future version adds any customer-facing text, clear it through search_product_catalog / Epiphan Knowledge search and check_my_copy first.

<outcome_sidecar>

{"ts":"...","skill":"team-coaching-rollup","version":"1.0.0",
 "running_user":"[email]","period":"[last-7-days|daily|custom]","status":"[success|partial|error]","runtime_ms":[n],
 "reps":[{"email":"[...]","owner_id":"[id]","nooks_id":"[id]","dials":[n],"connects":[n],"connect_rate":[f],
          "demos":[n],"callbacks_owed":[n],"top_disposition":"[live name]","flags":["..."],"focus":"[...]",
          "hubspot_nooks_gap":"[none|note]"}],
 "team":{"dials":[n],"connects":[n],"connect_rate":[f],"demos":[n],"median_connect_rate":[f]},
 "source_of_truth":{"volume_connects":"hubspot_rep_activity","disposition_mix":"nooks"},
 "slack_dm_sent_to_runner":[bool],"writes":"read-only-except-slack-dm","geo_scope":"north_america_usa_canada",
 "error":null,"session_id":"[YYYY-MM-DD]"}

</outcome_sidecar>

<success_criteria>

  • Every rep scoped by their own HubSpot owner id and Nooks user id; reps never blended
  • HubSpot rep_activity (query_dataset) used as the source of truth for dials and connects; Nooks (listCalls / listCallDispositions) used for the disposition mix and callbacks
  • HubSpot-vs-Nooks gaps flagged, not silently resolved
  • Scoreboard with per-rep metrics, team totals, and team median; demos surfaced as the headline outcome
  • Coaching flags raised with a concrete, humane focus per rep (at most two or three), strengths named too
  • Default period last 7 days; daily supported
  • Read-only everywhere except the optional Slack DM to the running user; no HubSpot, CRM, or Nooks record written
  • HTML rollup and sidecar written </success_criteria>
  • period [last-7-days]: last-7-days, daily, or a custom range.
  • dials_target [50/day]: the team daily dial target used for pace.
  • connect_band [8-12%]: the team connect-rate band used for the low-connect flag.
  • slack_summary [off]: DM the running user the team summary.
  • output [html+chat]: html+chat, or chat-only.

Changelog

v1.0.0 (initial):

  • Team activity and coaching rollup for the Epiphan SDR team, built for Tim and Victor. Pulls each rep's dials, connects, connect rate, disposition mix, callbacks, and demos booked for the period, scoped by each rep's HubSpot owner id and Nooks user id.
  • Reconciles the two sources the way the morning brief does: HubSpot rep_activity (query_dataset) is the source of truth for volume and connects, Nooks (listCalls / listCallDispositions) for the disposition mix and callbacks; HubSpot-vs-Nooks gaps are flagged.
  • Computes a scoreboard with team totals and median, and surfaces coaching flags (low connect rate, weak disposition hygiene, no demos, heavy no-answer) with a concrete focus per rep, naming strengths too.
  • Default period last 7 days, daily supported. GATE D North-America framing; no customer-facing copy so B and C do not apply. Read-only everywhere except an optional Slack DM to the running user. Delivers a chat scoreboard, an HTML rollup, and the sidecar.