dvcrn ClawdHub Contributor
Adds contribution capabilities to your agent. Help grow the ClawdHub ecosystem by analyzing skills, reporting bugs, and sharing useful skill combinations.
Modes
| Mode | What it does |
|---|---|
| passive | Bug reports and recipes only (safe default) |
| active | Adds auto-scout: analyze unknown skills locally and generate reports |
| full | All above plus opt-in telemetry |
Set mode in config/default.json or via CLAWDHUB_CONTRIB_MODE env var.
Features
Auto-Scout (active/full mode)
Analyze a local skill directory and produce a structured quality/security report:
bash scripts/scout.sh /path/to/skill-directory
Output: JSON report with dependency info, quality score, and security flags. Fully offline — no network access, pure static analysis.
Bug Reporting (all modes)
Report a skill failure with sanitized system context:
bash scripts/report-bug.sh <skill-slug> <error-message> [context]
Output: JSON bug report ready for API submission. Collects OS and node version but never leaks hostname, IP, or username.
Recipes (all modes)
Share a useful combination of skills that solved a task:
bash scripts/submit-recipe.sh <task-description> <skill1> [skill2] [skill3] ...
Output: JSON recipe ready for API submission.
Configuration
Edit config/default.json:
{
"mode": "passive",
"telemetry": false,
"autoScout": false,
"bugReports": true,
"recipes": true
}
| Key | Type | Description |
|---|---|---|
mode |
string | passive, active, or full |
telemetry |
bool | Opt-in anonymous usage stats (full mode only) |
autoScout |
bool | Auto-scan skills on encounter (active/full mode) |
bugReports |
bool | Enable bug report generation |
recipes |
bool | Enable recipe submission |
Commands Summary
| Command | Mode Required | Description |
|---|---|---|
scripts/scout.sh <dir> |
active+ | Analyze a skill directory |
scripts/report-bug.sh <slug> <msg> [ctx] |
any | Generate bug report |
scripts/submit-recipe.sh <task> <skills...> |
any | Generate recipe |
Security
- Scripts never access the network
- No credentials, IPs, hostnames, or usernames are collected
- All output is sanitized JSON to stdout
- Scout performs read-only static analysis