Slack User Context

Slack OAuth Scopes Used

| Scope | Purpose | |-------|---------| | users:read | View workspace members and their basic info | | users.profile:read | View detailed profiles (title, department, timezone) | | users:write | Set NetClaw's own presence status | | dnd:read | Check Do Not Disturb before escalating | | im:read | View DM channel info | | im:history | Read DM context for personalized responses | | channels:read | Know which channels exist and their purposes | | search:read.public | Search public channels for relevant discussions |

User-Aware Behaviors

1. Escalation Routing

Before escalating an alert to a specific person:

1. Check DND status — Is the engineer in Do Not Disturb mode?
2. Check presence — Are they active, away, or offline?
3. Check timezone — Is it their working hours?
4. Check profile title — Are they the right role for this issue?

Escalation Decision Matrix:

| DND | Presence | Severity | Action | |-----|----------|----------|--------| | Off | Active | Any | Mention directly | | Off | Away | P1/P2 | Mention (they'll see notification) | | Off | Away | P3/P4 | Post to channel, don't mention | | On | Any | P1 | Mention anyway (override DND for P1) | | On | Any | P2 | Try next person in rotation | | On | Any | P3/P4 | Post to channel only |

2. Personalized Response Depth

Adjust response complexity based on who's asking:

From user profile title/department:

| Title Contains | Response Style | |----------------|---------------| | "Network Engineer", "CCIE" | Full technical detail, CLI output, protocol specifics | | "NOC", "Operations" | Actionable steps, severity, impact summary | | "Manager", "Director" | Executive summary, business impact, timeline | | "Security", "InfoSec" | Focus on security implications, CVEs, compliance | | "Developer", "DevOps" | API perspectives, connectivity impact, service mapping | | "Help Desk", "Support" | Simple status, ETA, who to contact |

3. Timezone-Aware Scheduling

When scheduling reports or maintenance windows:

• Read the requesting user's timezone from their Slack profile
• Present times in THEIR timezone (not just UTC)
• Warn if proposed maintenance window is outside their business hours
• Suggest alternatives that work across timezones for multi-region teams

4. Context from Previous Conversations

When a user asks a follow-up question:

• Check recent DM and channel history for context
• Reference previous reports or investigations
• Don't re-explain concepts they've already discussed
• Build on previous analysis rather than starting fresh

Team Awareness Features

On-Call Roster Awareness

When NetClaw needs to escalate:

:telephone_receiver: *Escalation Attempt*
Checking team availability...

• @engineer1 — :green_circle: Active (Network Engineer) — *SELECTED*
• @engineer2 — :red_circle: DND until 08:00 EST (Senior NetEng)
• @engineer3 — :yellow_circle: Away (NOC Analyst) — last active 2h ago
• @manager1 — :green_circle: Active (Net Ops Manager) — CC'd on P1

Escalating to @engineer1...

Workload Context

If NetClaw sees a user is already active in multiple incident threads:

:information_source: Note: @engineer1 is currently active in 2 other incident threads.
Consider routing to @engineer3 if this is non-urgent.

Handoff Between Shifts

When a shift change is approaching:

:arrows_counterclockwise: *Shift Handoff Summary for @nightshift_engineer*

*Active Items:*
1. :warning: R2 memory trending high (78% → 82% over 8h) — monitoring
2. :wrench: SW1 firmware upgrade scheduled for 02:00 UTC — prep done
3. :white_check_mark: R1 Loopback99 change completed and verified

*Pending:*
4. :hourglass: Waiting on ISP ticket #45678 for R3 circuit
5. :clipboard: Security audit scheduled for SW2 tomorrow

_Full context in #netclaw-reports thread →_

NetClaw Presence Management

Set NetClaw's own presence to reflect its status:

| NetClaw Status | Slack Presence | Meaning | |----------------|---------------|---------| | Idle | :green_circle: Active | Ready for requests | | Running health check | :green_circle: Active | Processing (responsive) | | Running fleet-wide ops | :yellow_circle: Away | Batch processing (may be slow) | | Error / unreachable MCP | :red_circle: DND | Degraded — some skills unavailable |

Search for Context

Before answering questions about the network, search public channels for recent discussions:

• Search for device names mentioned in recent messages
• Find previous troubleshooting threads about the same device
• Reference prior decisions about the network architecture
• Link to relevant existing threads instead of duplicating information

Privacy and Boundaries

• Never expose user DND details, exact schedules, or personal info to other users
• Only use profile information to improve routing and response quality
• Don't track user activity patterns beyond immediate operational needs
• Respect when users don't want to be disturbed — follow the escalation matrix
• Be transparent about what context you're using when personalizing responses

Integration with Other Skills

| Skill | User Context Enhancement | |-------|-------------------------| | slack-network-alerts | Route to active, available engineers | | slack-incident-workflow | IC assignment based on availability | | slack-report-delivery | Adjust detail level per audience | | All pyATS skills | Include requestor context in GAIT trail |