gilles/StrixCamDB
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

Add StrixCamDB-Add and StrixCamDB-New-Protocol-Or-Placeholders skills

Browse Source
This commit is contained in:
eduard256
2026-03-23 16:45:01 +00:00
parent 539994b1a2
commit 57eda87fdf
2 changed files with 368 additions and 0 deletions
Download Patch File Download Diff File Expand all files Collapse all files
.claude/skills/StrixCamDB-Add/SKILL.md
+213
View File
@@ -0,0 +1,213 @@
---
name: StrixCamDB-Add
description: Add new brands, camera models, or stream URLs to StrixCamDB. Use when user wants to add camera data to the database, mentions adding a brand/model/URL, or provides camera stream information.
argument-hint: "[description of what to add]"
disable-model-invocation: true
---
# StrixCamDB-Add
You are editing the StrixCamDB camera database. Follow the rules below exactly.
Chat with the user in Russian. All data in JSON files -- in English.
### CRITICAL RULE -- READ-ONLY FOR EXISTING DATA
You can ONLY ADD new data: new brands, new streams, new models. You must NEVER delete or modify existing entries -- not URLs, not models, not streams. Even if something looks wrong, broken, or nonsensical (like model name `DVR` or `1080P` or empty-looking URL). These entries may be valid for specific cameras.
If you see something suspicious -- tell the user. But don't touch it without explicit confirmation.
---
## STEP 1: Determine the operation
If the user provided details in the command arguments -- parse them and determine what to do automatically. If not -- ask using AskUserQuestion:
**What do you want to do?**
1. **Add new brand** -- create a new JSON file for a brand that doesn't exist yet
2. **Add stream URL** -- add a new stream entry to an existing brand
3. **Add model to stream** -- add a model name to an existing stream's `models` array
---
## STEP 2: Collect required information
### For "Add new brand"
Ask (or parse from input):
- Brand name (human-readable, e.g. `Hikvision`)
- At least one stream URL with protocol and port
Brand ID is auto-generated from brand name: lowercase, spaces to hyphens, special chars removed.
### For "Add stream URL"
Ask (or parse from input):
- Brand name or brand_id (search in `brands/` directory if unclear)
- URL path (e.g. `/Streaming/Channels/101`)
- Protocol (`rtsp`, `http`, `https`, `rtsps`, `rtmp`, `mms`, `bubble`, `rtp`)
- Port number (e.g. `554`, `80`. Use `0` if unknown)
- Which models this stream works for (list of model names, or `*` for all)
### For "Add model to stream"
Ask (or parse from input):
- Brand name or brand_id
- Model name to add
- Which stream(s) to add it to (show existing streams, let user pick)
---
## STEP 3: Validate before writing
### Brand file structure
Every brand file is `brands/{brand_id}.json`:
```json
{
"version": 2,
"brand": "Brand Name",
"brand_id": "brand-name",
"streams": [
{
"id": "brand-name-1",
"url": "/path/to/stream",
"protocol": "rtsp",
"port": 554,
"models": ["Model-A", "Model-B"],
"notes": "Optional notes"
}
]
}
```
### Root fields
| Field | Type | Required | Rules |
|-------|------|----------|-------|
| `version` | int | yes | Always `2` |
| `brand` | string | yes | Human-readable name, capitalized properly |
| `brand_id` | string | yes | Lowercase, hyphens only, must match filename |
| `streams` | array | yes | At least one stream |
### Stream fields
| Field | Type | Required | Rules |
|-------|------|----------|-------|
| `id` | string | yes | Format: `{brand_id}-{N}` where N is sequential. Must be unique within file |
| `url` | string | yes | URL path only (no protocol://host:port prefix). Can contain placeholders |
| `protocol` | string | yes | One of: `rtsp`, `http`, `https`, `rtsps`, `rtmp`, `mms`, `bubble`, `rtp` |
| `port` | int | yes | 0-65535. Use `0` if unknown (means "use default for protocol") |
| `models` | array | yes | Non-empty. Use `["*"]` if stream works for all models of this brand |
| `notes` | string | no | Only add if genuinely useful context exists |
### Supported URL placeholders
| Placeholder | Description | Example |
|-------------|-------------|---------|
| `[CHANNEL]` | Channel number, 0-based | `0`, `1`, `2` |
| `[CHANNEL+1]` | Channel number, 1-based | `1`, `2`, `3` |
| `[USERNAME]` | Login username | `admin` |
| `[PASSWORD]` | Login password | `12345` |
| `[WIDTH]` | Video width | `1920` |
| `[HEIGHT]` | Video height | `1080` |
| `[IP]` | Camera IP address | `192.168.1.100` |
| `[PORT]` | Port number | `554` |
| `[AUTH]` | Base64-encoded `username:password` | |
Alternative forms: `[USER]`, `[PASS]`, `[PWD]`, `[PASWORD]`, `{CHANNEL}`, `{channel+1}` -- all supported.
### Validation rules
1. `brand_id` must match the filename (without `.json`)
2. Every `id` must be unique within the file
3. No duplicate streams -- same `protocol:port:url` combination must not appear twice
4. `url` must not be empty
5. `models` must not be empty
6. `port` must be integer 0-65535
7. `protocol` must be a non-empty string
8. If brand file already exists -- read it first, don't overwrite
---
## STEP 4: Write changes
### Adding a new brand
1. Verify file `brands/{brand_id}.json` does NOT exist
2. Create the file with `version: 2`, brand info, and streams
3. Stream IDs start from `{brand_id}-1`
### Adding a stream to existing brand
1. Read existing `brands/{brand_id}.json`
2. Find the highest existing stream ID number (e.g. if last is `dahua-47`, next is `dahua-48`)
3. Append new stream to the `streams` array
4. Write file back with `indent=2`, `ensure_ascii=False`, trailing newline
### Adding a model to existing stream
1. Read existing `brands/{brand_id}.json`
2. Find the target stream by ID or URL
3. Add model name to the `models` array (if not already present)
4. Write file back
### JSON formatting rules
- 2-space indent
- No trailing commas
- `ensure_ascii: false` (preserve Unicode)
- Single trailing newline at end of file
- Keys order in root: `version`, `brand`, `brand_id`, `streams`
- Keys order in stream: `id`, `url`, `protocol`, `port`, `models`, `notes`
---
## STEP 5: Verify
After writing:
1. Run `python3 scripts/validate.py` to check for errors
2. If validation fails -- fix the issue immediately
3. Show the user what was changed (brief summary)
### Consistency check -- IMPORTANT
Before finishing, check for problems and report them to the user immediately:
- Duplicate streams (same `protocol:port:url`) within the brand file
- Model names that look like categories, not real models (e.g. `DVR`, `PTZ`, `1080P`, `Other`) -- warn the user but don't block
- Unknown protocol values not listed in `schemas/brand.schema.json` -- warn the user, suggest running `/StrixCamDB-New-Protocol-Or-Placeholders`
- Placeholder in URL that is not documented in README.md -- warn the user
- Any other inconsistencies between schemas, README, and actual data -- report immediately
---
## STEP 6: Commit and version
Ask the user using AskUserQuestion:
**Question:** "Commit changes?"
- **Commit only** -- commit to main, CI will update latest release automatically
- **Commit + fix version** -- commit and create a version tag (ask for version, e.g. `v0.2.0`)
- **Don't commit** -- leave changes uncommitted
### If committing:
1. Stage changed files: `git add brands/{brand_id}.json`
2. Commit with message in AlexxIT style:
- New brand: `Add {brand} camera database`
- New stream: `Add stream URL for {brand}`
- New model: `Add {model} to {brand}`
3. Push to main: `git push origin main`
### If fixing version:
1. After commit and push, create tag: `git tag v{X.Y.Z}`
2. Push tag: `git push origin v{X.Y.Z}`
3. Confirm that CI will create a versioned release
### Commit message rules -- ABSOLUTE:
- One line, imperative verb, no period at the end
- No `feat:`, `fix:`, `chore:` prefixes
- No emoji, no "Co-Authored-By", no mention of AI/Claude
- Examples: `Add Reolink camera database`, `Add RTSP stream for TP-Link Tapo`, `Add DS-2CD2047 to Hikvision`
.claude/skills/StrixCamDB-New-Protocol-Or-Placeholders/SKILL.md
+155
View File
@@ -0,0 +1,155 @@
---
name: StrixCamDB-New-Protocol-Or-Placeholders
description: Add a new protocol or placeholder to StrixCamDB. Updates schemas, README, and skill definitions to support the new value across the entire project.
argument-hint: "[protocol name or placeholder]"
disable-model-invocation: true
---
# StrixCamDB-New-Protocol-Or-Placeholders
You are adding a new protocol or URL placeholder to the StrixCamDB project. This requires updating multiple files to keep everything in sync.
Chat with the user in Russian. All file content -- in English.
---
## STEP 1: Determine the operation
If the user provided details in the command arguments -- parse them. If not -- ask using AskUserQuestion:
**What do you want to add?**
1. **New protocol** -- e.g. `webrtc`, `onvif`, `p2p`
2. **New placeholder** -- e.g. `[STREAM]`, `[PROFILE]`
3. **Both** -- new protocol and new placeholder at the same time
---
## STEP 2: Collect information
### For new protocol
Ask (or parse from input):
- Protocol name (lowercase, e.g. `webrtc`)
- Short description (what it does, when it's used)
- Default port (if known, otherwise `0`)
### For new placeholder
Ask (or parse from input):
- Placeholder name in bracket format (e.g. `[STREAM]`)
- What it represents (e.g. "Stream profile index")
- Example value (e.g. `0`, `main`)
- Are there alternative forms? (e.g. `{stream}`, `[STREAM_ID]`)
---
## STEP 3: Update files
### 3A: Adding a new protocol
Update these 4 files:
**1. `schemas/brand.schema.json`**
Find the `protocol` field description in `$defs/stream/properties/protocol`. Add the new protocol to the description list.
Before:
```
"description": "Network protocol: rtsp, http, https, mms, rtmp, rtsps, bubble, rtp, or future protocols"
```
After (example adding `webrtc`):
```
"description": "Network protocol: rtsp, http, https, mms, rtmp, rtsps, bubble, rtp, webrtc, or future protocols"
```
**2. `schemas/preset.schema.json`**
Same change -- find `protocol` field description in `$defs/preset_stream/properties/protocol` and add the new protocol.
**3. `README.md`**
Find the protocol table in the "Database Format" section under "Stream fields". Add the new protocol to the `protocol` field description cell, keeping the same comma-separated format.
**4. `.claude/skills/StrixCamDB-Add/SKILL.md`**
Two places to update:
a) In the "For Add stream URL" section, find the protocol list:
```
- Protocol (`rtsp`, `http`, `https`, `rtsps`, `rtmp`, `mms`, `bubble`, `rtp`)
```
Add the new protocol.
b) In the "Stream fields" table, find the `protocol` row and add the new protocol to the description.
---
### 3B: Adding a new placeholder
Update these 3 files:
**1. `schemas/brand.schema.json`**
Find the `url` field description in `$defs/stream/properties/url`. Add the new placeholder to the list.
**2. `README.md`**
Find the "Placeholders" section. Add a new row to the table:
```markdown
| `[NEW_PLACEHOLDER]` | Description of what it does |
```
If there are alternative forms, add them to the note below the table.
**3. `.claude/skills/StrixCamDB-Add/SKILL.md`**
Find the "Supported URL placeholders" table. Add a new row with the placeholder, description, and example value.
If there are alternative forms, add them to the "Alternative forms" line below the table.
---
## STEP 4: Verify
1. Read each updated file and confirm the changes are correct
2. Run `python3 scripts/validate.py` to make sure nothing broke
3. Show the user a summary of all changes
### Consistency check -- IMPORTANT
Before finishing, compare ALL sources of truth and report any discrepancies to the user:
- Protocol list in `schemas/brand.schema.json` vs `schemas/preset.schema.json` vs `README.md` vs `StrixCamDB-Add/SKILL.md` -- must be identical
- Placeholder list in `schemas/brand.schema.json` vs `README.md` vs `StrixCamDB-Add/SKILL.md` -- must be identical
- If any file is out of sync (e.g. someone added a protocol to README but not to schema) -- fix it now and tell the user
- If `brands/*.json` contains protocol values not listed in schemas -- warn the user
---
## STEP 5: Commit
Ask the user using AskUserQuestion:
**Commit changes?**
- **Commit only** -- commit to main, CI updates latest release
- **Commit + fix version** -- commit and create a version tag (ask for version)
- **Don't commit** -- leave uncommitted
### If committing:
1. Stage all changed files
2. Commit message in AlexxIT style:
- Protocol: `Add {protocol} protocol support to database schema`
- Placeholder: `Add {placeholder} placeholder support`
- Both: `Add {protocol} protocol and {placeholder} placeholder support`
3. Push to main
### If fixing version:
1. After push, create tag: `git tag v{X.Y.Z}`
2. Push tag: `git push origin v{X.Y.Z}`
### Commit message rules -- ABSOLUTE:
- One line, imperative verb, no period at the end
- No `feat:`, `fix:`, `chore:` prefixes
- No emoji, no "Co-Authored-By", no mention of AI/Claude