Trending Skill Finder

Monitors ClawHub skill metrics and alerts when download growth exceeds 30%, downloads pass 50,000, stars pass 200, or new top 10 skills appear.

openclaw skills install @gabriel-kaufman/skill-surge-notifier

Skill Surge Notifier

Monitors ClawHub for skills gaining traction. Tracks download and star counts over time and alerts when a skill is surging.

A surge is triggered when any of these are true:

Requires Node.js 18+. No install needed — runs directly from the skill bundle.

Command	What it does
`node {baseDir}/cli.js fetch`	Show top 20 skills by downloads
`node {baseDir}/cli.js check`	Run surge detection, top movers, and update state
`node {baseDir}/cli.js status`	Last check, thresholds, notification status
`node {baseDir}/cli.js profile`	Show current agent profile
`node {baseDir}/cli.js profile set "description" "kw1,kw2"`	Set profile for relevance scoring
`node {baseDir}/cli.js config movers=5`	Set number of top movers shown
`node {baseDir}/cli.js config movers-off`	Disable top movers
`node {baseDir}/cli.js config growth=30 downloads=50000 stars=200`	Update surge thresholds

Every check run always shows top N movers (by download delta) regardless of thresholds, so there's always something to look at.

When a profile is set, surges are scored 0-10 for relevance and sorted accordingly — most relevant first.

State is stored in ~/.skill-surge-notifier/state.json. The first run seeds the baseline; growth % appears from the second run onward.
All output is printed to stdout. When used inside an agent, the agent must capture the CLI output to surface alerts in chat.
To run automatically, add to crontab (crontab -e):

bash

0 */4 * * * node {baseDir}/cli.js check >> ~/.skill-surge-notifier/surge.log 2>&1

These optional env vars override default paths and behavior:

Variable	Default	Description
`SURGE_DIR`	`~/.skill-surge-notifier`	Base directory for all state and config files
`STATE_PATH`	`$SURGE_DIR/state.json`	Path to the skill state file
`CONFIG_PATH`	`$SURGE_DIR/config.json`	Path to the config file
`SCHEDULED`	`false`	Set to `true` when running from a scheduler; adds a random 1–5 min delay before fetching to spread API load