Install
$ agentstack add skill-stephenrogan-augmented-csm-ra-stakeholder-change-detector ✓ scanned · ✓ verified — works with Claude Code, Cursor, and more.
Security review
✓ PassedNo issues found. Passed automated security review. · v0.1.0 How review works →
- ✓ Prompt-injection patterns
- ✓ Secret / credential exfiltration
- ✓ Dangerous shell & filesystem operations
- ✓ Untrusted network calls
- ✓ Known-malicious package signatures
What it can access
- ✓ Network access No
- ✓ Filesystem access No
- ✓ Shell / process execution No
- ✓ Environment & secrets No
- ✓ Dynamic code execution No
From automated source analysis of v0.1.0. “Used” means the capability is present in the source — more access means more to trust, not that it’s unsafe.
About
Stakeholder Change Detector
Monitors for stakeholder changes that could affect the customer relationship. Part of the Relationship Architecture pillar -- produces the highest-urgency signals in this pillar, particularly champion departure.
This is a detection and alerting skill. It identifies that a change happened. The CSM interprets the impact and decides the response. A champion departure detected in 48 hours instead of 4 weeks can be the difference between saving and losing the account.
When to Run
- Continuous: Monitors CRM contact updates and engagement pattern changes in real-time
- Scheduled: Weekly external scan (LinkedIn, if integration available) for role changes among key contacts
- Triggered: When a contact record changes in CRM or when an engagement pattern shifts suddenly
Core Execution Logic
Step 1: Monitor CRM-Based Signals (High Confidence)
| Signal | Detection Method | Confidence | Alert Urgency | |--------|-----------------|-----------|---------------| | Contact deactivated | CRM contact status changed to inactive or deleted | High | Immediate if Champion or Economic Buyer; same day for others | | Contact role changed | CRM title or role field updated | High | Same day -- assess whether the new role changes their influence | | New contact added | New contact record created on the account | Medium | This week -- classify and assess | | Contact reassigned | Contact moved to a different account or business unit | High | Same day -- equivalent to departure for this account | | Ownership field changed | Account's CSM or AE field changes (internal change, but affects the customer) | High | Immediate -- triggers lo-handoff-manager |
Step 2: Monitor Engagement-Based Signals (Medium Confidence)
| Signal | Detection Criteria | Confidence | Possible Explanations | |--------|-------------------|-----------|----------------------| | Sudden silence | A previously active contact (touchpoint every 2 weeks or more) has zero engagement for 30+ days | Medium | Departed, went on leave, disengaging, or simply busy. Investigate before assuming departure | | Email bounce | Outbound email returns a bounce (invalid address) or auto-reply indicating departure | High | Strong departure signal. Auto-replies with "I am no longer at [company]" are definitive | | Meeting delegate | A contact consistently sends a delegate to meetings they previously attended personally (2+ times) | Medium | May indicate disengagement, role change, or delegation due to seniority. Not necessarily departure | | New voice emerging | A previously unknown contact begins appearing in calls, emails, or ticket interactions without introduction | Low-Medium | May indicate a replacement for someone who left, a reorg, or natural team expansion. Investigate | | Response pattern change | A contact who previously responded within 24 hours now consistently takes 5+ days | Medium | May indicate disengagement, workload change, or deprioritisation of the vendor relationship |
Step 3: Monitor External Signals (Variable Confidence)
If LinkedIn integration or similar service is available:
| Signal | Detection Method | Confidence | Latency | |--------|-----------------|-----------|---------| | Role change (same company) | LinkedIn title update showing new role at the same employer | High | 1-4 weeks after actual change (people update LinkedIn on their own schedule) | | Company departure | LinkedIn shows new employer | High | 1-6 weeks after actual departure | | New hire at customer | Person with relevant title joins the customer company | Medium | 1-4 weeks after start date | | Reorg signal | Multiple contacts at the same company update titles simultaneously | Medium | Indicates restructuring. May affect reporting lines and decision-making authority |
Important: LinkedIn data is lagging. A champion may have left 3 weeks before updating their profile. Engagement-based signals (sudden silence, email bounce) often detect departures faster.
Step 4: Classify and Score Change Impact
For each detected change, assess the impact on the account:
| Change Type | Impact Assessment Factors | Severity Range | |-------------|--------------------------|---------------| | Champion departure | Account ARR, renewal proximity, relationship breadth (single-threaded?), replacement identified? | Critical (5) if single-threaded or near renewal. High (4) if well-threaded | | Economic buyer change | New buyer's disposition (known/unknown), budget cycle timing, prior relationship | High (4) if unknown disposition. Medium (3) if new buyer is already engaged | | Executive sponsor departure | Strategic account? Air cover needed for budget protection? | High (4) for strategic accounts. Medium (3) for standard | | Technical lead change | Active integrations or implementations in progress? | Medium (3) if active project. Low (2) if stable usage | | New hire (relevant role) | Potential new champion or economic buyer? Opportunity to build relationship? | Low (1) -- opportunity signal, not risk | | Reorg | Scope and relevance of restructuring to the product's stakeholder ecosystem | Variable -- depends on which roles and relationships are affected |
Step 5: Generate Alert
{
"alert_id": "sc-2026-0089",
"account_id": "string",
"account_name": "Acme Corp",
"change_type": "champion_departure",
"severity": 5,
"contact": {
"name": "Jane Smith",
"previous_role": "VP Product",
"stakeholder_type": "champion",
"detection_method": "linkedin_role_change",
"detection_date": "2026-03-08",
"estimated_departure_date": "2026-02-28",
"new_employer": "Competitor X",
"new_role": "VP Product"
},
"impact_assessment": {
"account_arr": 85000,
"days_to_renewal": 112,
"other_active_contacts": 2,
"single_threaded_risk": false,
"health_score": 72,
"health_trend": "declining",
"relationship_breadth": "moderate -- 2 other active contacts but no executive coverage"
},
"correlated_signals": [
"Usage declining for 3 weeks (bi-usage-monitor) -- may correlate with champion departure timeline",
"No executive sponsor engaged (ra-stakeholder-mapper)"
],
"recommended_actions": [
"Identify replacement champion or sponsor immediately",
"Reach out to remaining active contacts to confirm continuity and assess impact",
"Update stakeholder map in ra-stakeholder-mapper",
"Brief CSM manager -- champion departure at a mid-sized account approaching renewal",
"Monitor for further engagement decline in the next 14 days"
],
"urgency": "immediate",
"routed_to": ["CSM: Jane Doe", "Manager: Mike Ross"]
}
Step 6: Track Change Lifecycle
Each stakeholder change follows a lifecycle:
| Status | Definition | |--------|-----------| | Detected | Change signal identified | | Confirmed | CSM or data source confirms the change is real (not a data error) | | Impact assessed | CSM has evaluated the effect on the relationship | | Response in progress | CSM is actively addressing the change (building new relationship, securing replacement) | | Stabilised | New stakeholder landscape is established and relationships are active |
Track time from detection to stabilisation. This measures how quickly the CS team responds to relationship disruption.
Handoff to Human
| Decision Point | Context Provided | What the Human Decides | |---------------|-----------------|----------------------| | Change confirmation | Signal evidence, detection method, confidence level | Whether the change is real or a false positive. Some signals (email bounce) are definitive; others (sudden silence) need investigation | | Impact assessment | Stakeholder role, account context, renewal proximity, relationship breadth | How serious this is for this specific account. A champion departure at a well-threaded account is concerning but manageable. At a single-threaded account near renewal, it is an emergency | | Response strategy | Recommended actions, remaining contacts, account brief | How to respond -- who to reach out to, what to say, whether to escalate to leadership, whether executive engagement is needed | | New contact engagement | New hire detection, role, account context | Whether and how to engage the new person. Timing and approach depend on the relationship with remaining contacts and the new person's likely disposition | | Opportunity assessment | New hire or reorg signals | Whether a change is a threat, an opportunity, or neutral. A new CTO might be a chance to re-engage at the executive level if the previous one was unresponsive |
Confidence and Limitations
- High confidence for CRM-based detection (contact deactivated, role changed) -- structured data events with defined triggers
- Medium confidence for engagement-based detection (sudden silence, meeting delegates, response delays) -- behavioural patterns have multiple possible explanations. The skill presents the signal; the CSM investigates the cause
- Low-Medium confidence for LinkedIn detection -- depends on integration availability, data freshness (people update profiles on their own schedule), and matching accuracy (common names, incomplete profiles)
- Cannot detect internal reorganisations that do not result in CRM contact record changes or LinkedIn updates. Reorgs may be invisible until their effects manifest in engagement patterns
- Cannot assess the incoming person's disposition toward the product. A new VP might be a champion or a sceptic -- the classification only becomes clear through engagement
- Cannot detect departures that happen without any digital signal (contact leaves but email remains active during notice period, LinkedIn is not updated)
- Champion departure to a competitor (as in the example above) is a double signal: loss of the champion AND potential competitive intelligence gained by the competitor
Dependencies
Required:
- CRM API (contact records, activity data, field change events)
- ra-stakeholder-mapper (stakeholder classifications and coverage context)
Strongly recommended:
- Email integration (bounce detection, auto-reply detection)
- Calendar integration (meeting delegate pattern detection)
- ra-engagement-tracker (engagement pattern baselines for silence detection)
- bi-risk-detector (champion departure feeds as a risk signal -- the highest-severity signal in the system)
Optional but high-value:
- LinkedIn integration or stakeholder monitoring service (for external detection)
Downstream consumers:
- bi-risk-detector (champion departure and executive change as risk signals)
- ra-stakeholder-mapper (triggers stakeholder map update)
- ic-internal-notifier (routes change alerts to appropriate recipients)
- lo-renewal-manager (stakeholder change near renewal amplifies renewal risk)
References
references/change-detection-methods.md-- Detection method details, confidence calibration, and false positive reduction strategies
Source & license
This open-source skill is cataloged on AgentStack and links to its original source — we do not rehost the code.
- Author: stephenrogan
- Source: stephenrogan/augmented-csm
- License: MIT
Install and usage instructions live in the source repository linked above.
Reviews
No reviews yet — be the first.
Write a review
Versions
- v0.1.0 Imported from the upstream source.