octavehq/deal-coach

/octave:deal-coach — Methodology Coaching

An interactive coaching skill built around the Resonate → Elevate → Compel sales methodology. Choose your output mode — role play, coaching microsite, coaching deck, or interactive quiz — and get coaching grounded in deal context AND your actual GTM messaging.

Three Stages: | Stage | Focus | |-------|-------| | Resonate | Understand and resonate with the buyer | | Elevate | Confirm the fit and elevate the opportunity | | Compel | Deliver the value and compel the buyer to action |

Three Coaching Outputs per Stage: | Field | Type | Purpose | |-------|------|---------| | Buyer Mindset | String | Where the buyer's head is — psychology, fears, motivations | | Value Propositions | Array | Which value props to deploy and why they fit this stage | | Talking Points | Array | Specific things to say, grounded in deal context |

This skill reads five coaching reference files at runtime:

• references/frameworks.md — Resonate / Elevate / Compel: Buyer Mindset + Value Props + Talking Points
• references/coaching-agents.md — 3 coaching agent personas + 2 cross-stage + scoring rubrics
• references/stage-mapping.md — Buyer's Journey → coaching stage mapping + inference rules
• references/html-templates.md — HTML section/slide templates per output mode
• references/messaging-narratives.md — How coaching grounds in GTM context

If reference files are not found, fall back to general coaching methodology and note the limitation.

How this differs from /octave:pipeline:

• /octave:pipeline gives live-deal strategy — diagnosis and next steps to move a specific deal forward
• /octave:deal-coach produces practice and coaching materials organized around the three coaching stages

How this differs from /octave:train:

• /octave:train is generic sales training (objection handling, personas, product knowledge)
• /octave:deal-coach is methodology-specific — every output is structured around Resonate/Elevate/Compel, scored against coaching rubrics, and coached by stage-specific agents

How this differs from /octave:meeting-prep:

• /octave:meeting-prep produces a strategic battle plan for an upcoming meeting
• /octave:deal-coach produces training and coaching materials organized around the three coaching stages

On-brand styling — use a brand kit if one exists

Before generating HTML output, decide whose brand this coaching asset should match (usually the target company; sometimes your own company), then follow brand-kit-usage.md: check for a cached kit, offer to use it, offer to capture one if missing, and fall back to a style preset if declined. The brand kit is the strongest styling signal — prefer it over generic --style presets.

Review pass (runs by default)

After generating, run the review pass by default — don't wait to be asked. In interactive mode, tell the user at intake that you'll review before finishing (recommended) and that they can opt out with --skip-review or "skip review". Follow review-pass.md for the preflight and visual pass. When generating, follow the output rules in presentation-principles.md — label every value, no tool names in the output, confirmed vs hypothesized, lean and deal-specific.

Usage

/octave:deal-coach
/octave:deal-coach [company domain or name]
/octave:deal-coach --mode [roleplay|microsite|deck|quiz]
/octave:deal-coach --stage [resonate|elevate|compel]
/octave:deal-coach [domain] --mode roleplay --stage elevate

Examples

/octave:deal-coach
/octave:deal-coach acme.com
/octave:deal-coach --mode roleplay
/octave:deal-coach acme.com --mode microsite --stage compel
/octave:deal-coach --mode quiz --stage resonate
/octave:deal-coach acme.com --mode deck --stage elevate

Instructions

Follow these steps precisely. Do not skip or reorder them.

---

Step 1: Choose Output Type (CT-1)

If the user specified --mode, use that. Otherwise, ask:

AskUserQuestion({
  questions: [{
    question: "What kind of coaching output do you want?",
    header: "Output Mode",
    options: [
      {
        label: "Role Play",
        description: "Practice a coaching-backed conversation scored against Buyer Mindset, Value Props, and Talking Points. 8-12 exchanges, then a scorecard."
      },
      {
        label: "Coaching Microsite",
        description: "Self-contained HTML coaching page organized around Buyer Mindset / Value Propositions / Talking Points — with deal context, priority actions, and stage-grounded messaging."
      },
      {
        label: "Coaching Deck",
        description: "Slide presentation walking through the coaching framework for a specific deal, with stage-specific content and talk tracks."
      },
      {
        label: "Interactive Quiz",
        description: "Test coaching methodology knowledge with deal-grounded scenarios. Scoring breaks down by Resonate/Elevate/Compel."
      }
    ],
    multiSelect: false
  }]
})

Store the selected mode for routing in Step 5.

---

Step 2: Identify Deal Context (CT-2)

Determine if coaching should be grounded in a specific deal or run in generic/practice mode.

If the user already provided a domain, name, or email, skip to 2b.

2a. Ask for deal context

AskUserQuestion({
  questions: [{
    question: "Should this be grounded in a specific deal, or do you want generic practice?",
    header: "Deal Context",
    options: [
      {
        label: "Specific Deal",
        description: "Ground coaching in a real account — uses CRM data, findings, events, and the matched Motion ICP narrative to make coaching deal-specific."
      },
      {
        label: "Generic Practice",
        description: "Practice the methodology with library-level data only (personas, Motion ICP narrative, proof points). No specific account context."
      }
    ],
    multiSelect: false
  }]
})

If Generic Practice: Skip to Step 2c.

If Specific Deal: Ask for the company domain or contact email (e.g., acme.com or [email protected]).

2b. Gather deal-specific context

Run these MCP tool calls to build a complete picture. Run independent calls in parallel:

Company & CRM Context (parallel):

enrich_company({ companyDomain: "<domain>" })
find_crm_records({ companyDomain: "<domain>" })
find_crm_activities({ companyDomain: "<domain>" })
generate_crm_context({ companyDomain: "<domain>" })

Library Context (parallel):

search_knowledge_base({ query: "<company name> OR <industry>" })
list_findings({
  query: "objections pain points competitor mentions next steps",
  eventFilters: { companyDomains: ["<domain>"] }
})
list_events({
  filters: { companyDomains: ["<domain>"] }
})
list_entities({ entityType: "persona" })
list_entities({ entityType: "competitor" })
list_entities({ entityType: "proof_point" })

Motion ICP (after enrichment / persona / segment inference):

list_motions()
list_motion_icps({ motionOId: "<motion_oId>" })
find_motion_icp({ motionIcpOId: "<matched_motion_icp_oId>", includeLearnings: true })

If a Custom Motion Playbook (Thematic / Milestone / Account / Competitive) applies, also fetch its narrative:

list_motion_playbooks({ motionOId: "<motion_oId>" })
get_motion_playbook({ motionPlaybookOId: "<custom_motion_playbook_oId>" })

If any tool call fails, note the gap and continue. Coach with available data and flag missing context.

2c. Generic practice mode context

For generic mode, gather library-level data only:

list_entities({ entityType: "persona" })
list_entities({ entityType: "competitor" })
list_entities({ entityType: "proof_point" })

Then load the Default Motion Playbook narrative for the most relevant Motion:

list_motions()
list_motion_icps({ motionOId: "<motion_oId>" })
find_motion_icp({ motionIcpOId: "<representative_motion_icp_oId>", includeLearnings: true })

2d. Load coaching reference files

Read the references needed for the inferred or selected stage:

Read: references/frameworks.md
Read: references/stage-mapping.md
Read: references/coaching-agents.md
Read: references/messaging-narratives.md

If generating HTML output (microsite or deck), also read:

Read: references/html-templates.md

---

Step 3: Infer Coaching Stage + User Override (CT-3)

If the user specified --stage, use that and skip to Step 3b.

3a. Infer stage from signals

Use the weighted inference algorithm from references/stage-mapping.md:

| Signal | Weight | Source | |--------|--------|--------| | CRM deal stage | 40% | find_crm_records → map to Resonate/Elevate/Compel | | Conversation findings | 30% | list_findings — pain points (→Resonate), competitor mentions (→Elevate), ROI/budget (→Compel) | | Deal activity patterns | 20% | list_events — discovery call (→Resonate), demo (→Elevate), proposal (→Compel) | | Time in stage | 10% | Days in current stage vs. expectation |

For generic practice mode: Skip inference. Let the user choose a stage:

AskUserQuestion({
  questions: [{
    question: "Which coaching stage do you want to practice?",
    header: "Stage",
    options: [
      { label: "Resonate", description: "Understand and resonate with the buyer — discovery principles, building trust through understanding" },
      { label: "Elevate", description: "Confirm the fit and elevate the opportunity — disrupt status quo, differentiate on value, build credibility" },
      { label: "Compel", description: "Deliver the value and compel the buyer to action — business case, Why Now, champion enablement" }
    ],
    multiSelect: false
  }]
})

3b. Present inference and allow override

Confidence calibration: CRM absence is a data hygiene issue, not a deal health signal. If CRM data is missing but activity signals are strong, redistribute the CRM weight across other signals.

Present the inference:

STAGE INFERENCE
===============
Stage: [Resonate / Elevate / Compel]
Confidence: [High / Medium / Low]
Buyer's Journey Phase: [Phase Name]

EVIDENCE
--------
CRM Stage (40%): "[Stage Name]" → maps to [Stage]  [or "No CRM record — data gap, not deal gap"]
Findings (30%): [Key signals]
Activities (20%): [Key activities]
Time (10%): [Assessment]

Then confirm with the user: proceed with the inferred stage, override and pick a different one, or see all three stages with descriptions before choosing (same list as generic mode above).

3c. Stall detection

If time-in-stage signals indicate a stalled deal (>2x expected time), flag it — but don't re-implement deal rescue here:

STALL DETECTION
===============
This deal appears stalled at [Journey Phase].
Time in stage: [X] days (expected: [Y] days)

Root Cause Hypothesis: [Stage] gap
- [Explanation of why this stage gap is likely the root cause]
- [Specific evidence from findings/activities]

For a re-engagement strategy and next steps on this deal, run:
/octave:pipeline stalled [domain]

For coaching, I'll focus on [Root Cause Stage], not the nominal CRM stage.

Route coaching to the root cause stage's coaching agent, not the nominal stage. Deal-rescue strategy itself belongs to /octave:pipeline.

---

Step 4: Route to Coaching Agent (CT-4)

Based on the confirmed stage, activate the appropriate coaching agent from references/coaching-agents.md:

| Stage | Coaching Agent | Focus | |-------|---------------|-------| | Resonate | Resonance Coach | Discovery principles (wide, deep, high), trust building | | Elevate | Elevation Coach | Case for Change, Value Framing, differentiated value, proof points | | Compel | Compel Coach | Business case building, Why Now Case, champion enablement |

Cross-stage agents (available as supplements):

• Negotiation Strategist — Available when stage is Compel and negotiation dynamics surface
• Objection Handler — Available at any stage when objections surface in findings

Load the agent's persona, coaching criteria, scoring rubric, and grounding instructions from the reference file. Use the grounding map from references/messaging-narratives.md to connect the agent's outputs to the seller's Octave library data.

---

Step 5: Generate Output (CT-5)

Branch based on the output mode selected in Step 1. Each mode has its own reference file with the full flow — read it and follow it:

| Mode | Reference | What it covers | |------|-----------|----------------| | Role Play | mode-roleplay.md | Scenario setup, stage-specific buyer psychology, running the conversation, coaching scorecard (RP-1 → RP-4) | | Coaching Microsite | mode-microsite.md | Style selection, content outline, HTML generation with grounding rules and density limits (MS-1 → MS-3) | | Coaching Deck | mode-deck.md | Style selection, stage-specific slide outlines, HTML deck generation on the deck skill's architecture (DK-1 → DK-3) | | Interactive Quiz | mode-quiz.md | Quiz type and length, question construction, running the quiz, results (QZ-1 → QZ-4) |

Role Play and Quiz run on the shared engine in roleplay-mechanics.md — the mode files layer the Resonate/Elevate/Compel specifics on top.

---

Step 6: Delivery + Next Actions (CT-6)

After delivering any output, offer iterations:

For Role Play and Quiz: Next actions are included in RP-4 and QZ-4 in the mode references.

For Microsite and Deck: After opening the HTML file, present:

Coaching [microsite/deck] generated and opened in your browser.

File: [file path]
Company: [Company Name or "Generic Practice"]
Stage: [Resonate / Elevate / Compel] — [Stage subLabel]
Coaching Agent: [Agent Name]
Style: [Preset Name]

Want to:
1. Practice this stage with a role play
2. Add MEDDPICC deal gap analysis [if not already included]
3. Move to the next stage ([Next Stage]: [subLabel])
4. Try a different output mode for this stage
5. Regenerate with a different style
6. Done

If the user picks option 3 (next stage), return to Step 3b with the next stage pre-selected and flow through Steps 4-5 again.

---

Output Directory

All HTML outputs go to .octave-deal-coach/ in the project root:

.octave-deal-coach/
  [company-kebab]-[stage]-[YYYY-MM-DD]/
    [company-kebab]-[stage].html            # Microsite
  [company-kebab]-deck-[stage]-[YYYY-MM-DD]/
    [company-kebab]-deck-[stage].html       # Deck

This directory should be in .gitignore.

---

MCP Tools Used

Research & Enrichment

| Tool | Purpose | |------|---------| | enrich_company | Company profile, industry, tech stack, strategic context | | find_crm_records | Deal stage, amount, close date, pipeline position | | find_crm_activities | Recent interactions — calls, emails, meetings | | generate_crm_context | AI-synthesized CRM narrative |

Library — Fetching

| Tool | Purpose | |------|---------| | list_motions | List all Motions in the workspace | | list_motion_playbooks | List Motion Playbooks (Default + Custom) under a Motion | | get_motion_playbook | Full details for a Motion Playbook | | list_motion_icps | List Motion ICP cells (persona × segment intersections) for a Motion | | find_motion_icp | Full Motion ICP cell narrative (Target ICP overview, Operating landscape, Strategic narrative, Pains and consequences, Benefits and impacts, Methodology, References) plus Learning Loop learnings | | get_entity | Individual entity details (persona, competitor, proof point) |

Library — Searching

| Tool | Purpose | |------|---------| | search_knowledge_base | Find matching guides and research | | search_resources | Find relevant resources (docs, presentations) | | list_entities | List personas, competitors, proof points, references | | list_findings | Objections, pain points, competitor mentions from conversations | | list_events | Deal history, stage changes, activity timeline |

Content Generation

| Tool | Purpose | |------|---------| | generate_content | Generate supporting content if needed |

---

Error Handling

No CRM data found: "I couldn't find CRM records for [domain]. I'll proceed with library-level data only. Stage inference will rely on findings and events. You can also manually select a stage."

No Motion ICP found: "No matching Motion ICP cell found. I'll use general coaching methodology without Motion-specific grounding. Talk tracks will reference the framework but won't include your specific Strategic narrative, Benefits and impacts, or Methodology content. Consider creating a Motion for this offering, or layering a Custom Motion Playbook (Thematic / Milestone / Account / Competitive) onto an existing Motion."

No findings/events: "No conversation findings or events found for [domain]. Stage inference will rely primarily on CRM stage. For more accurate coaching, ensure conversation data is synced to Octave."

Reference file not found: "Could not load [reference file]. Falling back to general coaching methodology. For full coaching, ensure reference files are in skills/deal-coach/references/."

Stage inference low confidence: "I'm not confident about the coaching stage — multiple stages scored similarly. I'd recommend selecting manually. Here are all options: [present all three stages]."

MCP connection failed: "Could not connect to Octave. Check your connection with /octave:workspace. This skill requires Octave MCP tools for deal context and library data."

HTML write failed: "Could not write the HTML file. Check that .octave-deal-coach/ is writable. Try: mkdir -p .octave-deal-coach"

---

Related Skills

• /octave:pipeline — Live-deal strategy: stalled-deal rescue, multi-threading, competitive threats, closing
• /octave:train — Generic sales training (role play, quiz, guided learning) without the coaching methodology
• /octave:meeting-prep — Strategic meeting battle plan as HTML
• /octave:deck — General-purpose slide deck builder
• /octave:enablement — Sales enablement materials (cheat sheets, objection guides)
• /octave:battlecard — Competitive intelligence and battlecards
• /octave:research — Account and person research