Compare commits

...

2 Commits

Author SHA1 Message Date
7bd2bb544a sync: auto-sync from DESKTOP-0O8A1RL at 2026-05-16 08:48:51
Author: Mike Swanson
Machine: DESKTOP-0O8A1RL
Timestamp: 2026-05-16 08:48:51
2026-05-16 08:48:55 -07:00
83e8e4498c feat: add /feature-request skill for Howard to submit GuruRMM roadmap requests 2026-05-16 08:48:44 -07:00
3 changed files with 175 additions and 0 deletions

View File

@@ -248,6 +248,7 @@ Vault structure: `infrastructure/`, `clients/`, `services/`, `projects/`, `msp-t
| `/create-spec` | Create app specification for AutoCoder | | `/create-spec` | Create app specification for AutoCoder |
| `/frontend-design` | Modern frontend design (auto-invoke after UI changes) | | `/frontend-design` | Modern frontend design (auto-invoke after UI changes) |
| `/remediation-tool` | M365 breach checks, tenant sweeps, gated remediation | | `/remediation-tool` | M365 breach checks, tenant sweeps, gated remediation |
| `/feature-request` | Howard submits a GuruRMM feature request — Claude classifies it and messages Mike |
--- ---

View File

@@ -0,0 +1,99 @@
Howard is submitting a GuruRMM feature request. Follow these steps exactly.
---
## Step 1 — Read context
Read `.claude/identity.json` to get the machine hostname (for session ID and Ollama endpoint).
Read `projects/msp-tools/guru-rmm/docs/FEATURE_ROADMAP.md`. Extract the top-level section headers
and a representative sample of subsections so you understand the existing structure.
Do NOT print the roadmap to the user.
---
## Step 2 — Classify the feature using Ollama
The feature request is: $ARGUMENTS
Determine the Ollama endpoint:
- If hostname is `DESKTOP-0O8A1RL`: use `http://localhost:11434`
- All other machines (including Howard's): use `http://100.92.127.64:11434`
Call Ollama with model `qwen3:14b` to classify the request. Ask it to return:
1. Which top-level roadmap section this belongs in (e.g., "Core Agent Features", "Server/API Features",
"Dashboard & UI", "Platform & Infrastructure", "Integrations", "Future Considerations")
2. Suggested subsection — match an existing one if it fits; propose a new name if it doesn't
3. Suggested priority (P1 = critical/blocking, P2 = important/near-term, P3 = nice-to-have)
4. A 2-3 sentence plain-English summary of what Howard is asking for
Use this prompt structure for the Ollama call:
```
You are helping classify a feature request for GuruRMM, a Remote Monitoring & Management tool
for MSPs. The roadmap sections are: Core Agent Features, Server/API Features, Dashboard & UI,
Platform & Infrastructure, Integrations, Future Considerations.
Feature request: <$ARGUMENTS>
Respond with JSON only: {"section": "...", "subsection": "...", "priority": "P1|P2|P3", "summary": "..."}
```
Parse the JSON response. If Ollama is unreachable, perform the classification yourself directly.
---
## Step 3 — Send coord message to Mike
Post a message to the coord API:
```bash
curl -s -X POST http://172.16.3.30:8001/api/coord/messages \
-H "Content-Type: application/json" \
-d '{
"from_session": "<HOSTNAME>/claude-main",
"to_session": "DESKTOP-0O8A1RL/claude-main",
"project_key": "gururmm",
"subject": "Feature Request from Howard: <brief title>",
"body": "<full body — see format below>"
}'
```
Where `<HOSTNAME>` comes from identity.json.
**Message body format:**
```
Howard submitted a feature request for GuruRMM.
FEATURE REQUEST
===============
<$ARGUMENTS verbatim>
PROPOSED PLACEMENT
==================
Section: <section from Ollama>
Subsection: <subsection from Ollama>
Priority: <P1/P2/P3>
SUMMARY
=======
<2-3 sentence summary from Ollama>
---
Submitted via /feature-request from <HOSTNAME> at <timestamp>
```
Send to `to_session: "DESKTOP-0O8A1RL/claude-main"` and also to `"Mikes-MacBook-Air/claude-main"`
so Mike sees it regardless of which machine he's on.
---
## Step 4 — Acknowledge to Howard
Tell Howard:
- His request was submitted to Mike
- Where you've suggested it fits in the roadmap (section + subsection)
- Suggested priority
- That Mike will review and decide if/how/when to build it
Keep it brief — one short paragraph. Do NOT say "I" did anything — say the request was sent.

75
tmp_roadmap_patch.py Normal file
View File

@@ -0,0 +1,75 @@
new_section = """
### Asset Location Tracking
MSP clients frequently ask "can you find this device if it's stolen?" — this feature makes the answer yes.
Primary use case is theft recovery and asset accountability, not real-time surveillance.
#### Approach by device type
**Laptops & desktops (Windows/macOS/Linux) — agent-based, no MDM required:**
Location is derived from visible WiFi networks (SSID + BSSID) sent to the Google or Microsoft
Geolocation API, which returns coordinates accurate to ~50-100m. Falls back to IP geolocation
(city-level) on wired-only or WiFi-disabled machines. No GPS hardware required.
**Phones & tablets — mobile agent app:**
A lightweight GuruRMM Agent app (iOS App Store / Google Play) reports background location and
can receive remote lock commands via standard APNs/FCM push — no Apple MDM vendor certificate
required. This covers the "find my stolen iPad" use case without the full MDM enrollment stack.
#### Agent (Windows/macOS/Linux)
- [ ] Collect visible WiFi networks (SSID, BSSID, signal strength) - P2
- [ ] Send to geolocation API (Google or Microsoft, configurable); store lat/lng + accuracy radius - P2
- [ ] IP geolocation fallback when WiFi scan returns empty - P2
- [ ] Location reported on each inventory checkin (not continuous polling) - P2
- [ ] Policy flag: location collection enabled/disabled per site or agent - P2
- [ ] On-demand location request (server pushes LocationRequest, agent responds immediately) - P2
- [ ] Platform WiFi scan implementations:
- [ ] Windows: `netsh wlan show networks mode=bssid` or `wlanapi` - P2
- [ ] macOS: CoreWLAN CWInterface.scanForNetworks - P2
- [ ] Linux: `iwlist scan` / nl80211 via netlink - P2
- [ ] Agent platform parity: all three platforms in same release per parity rule - P1
#### Server
- [ ] `agent_locations` table (agent_id, recorded_at, lat, lng, accuracy_meters, source: wifi|ip|gps, raw_wifi_scan JSON) - P2
- [ ] Location history — retain last N readings (configurable, default 30 days) - P2
- [ ] `POST /api/agents/:id/location` ingest endpoint - P2
- [ ] `GET /api/agents/:id/location` — current + history - P2
- [ ] `POST /api/agents/:id/location/request` — on-demand location pull - P2
- [ ] Fleet location endpoint: `GET /api/sites/:id/locations` (all agents for a site) - P2
#### Dashboard
- [ ] Location tab on Agent Detail page — map embed (Leaflet.js, OpenStreetMap tiles, no API key) with accuracy circle - P2
- [ ] Last known location + timestamp in Agent Overview card - P2
- [ ] Fleet map view — all online agents for a site pinned on one map - P3
- [ ] Location history timeline (scroll through past readings) - P3
- [ ] "Request Location Now" button — fires on-demand poll, updates map live - P2
- [ ] Geofence alerts — notify if agent appears outside defined boundary - P3
#### Mobile App (iOS + Android) — future phase
- [ ] Lightweight GuruRMM Agent app distributed via App Store / Google Play - P3
- [ ] Background location reporting (uses standard OS location APIs, not MDM) - P3
- [ ] Remote lock via APNs/FCM push notification (guided access on iOS, device admin on Android) - P3
- [ ] Enrollment via QR code / enrollment URL (same server, mobile-specific agent type) - P3
- [ ] No Apple MDM vendor certificate required — standard app distribution - P3
#### Out of scope (v1)
- Real-time continuous GPS tracking (battery/privacy concerns; on-demand + checkin is sufficient)
- Full MDM profile management for iOS/Android (separate MDM integration feature if ever needed)
- Wired-only machines where WiFi is disabled and IP geolocation accuracy is unacceptable
"""
with open('/home/guru/gururmm/docs/FEATURE_ROADMAP.md', 'r') as f:
content = f.read()
# Insert before the --- that separates Core Agent Features from Server/API Features
# That separator comes right after the tray security section ending at line ~183
marker = "- [ ] Optional PIN/password for sensitive actions - P3\n\n---\n\n## Server/API Features"
replacement = "- [ ] Optional PIN/password for sensitive actions - P3\n" + new_section + "\n---\n\n## Server/API Features"
if marker in content:
content = content.replace(marker, replacement, 1)
with open('/home/guru/gururmm/docs/FEATURE_ROADMAP.md', 'w') as f:
f.write(content)
print("inserted")
else:
print("MARKER NOT FOUND")