메뉴
Configurable regex-based alert system for detecting patterns in tool calls and session activity. Supports error triggers, content regex matching, token threshold triggers, and pattern detection with configurable actions.
기본적으로 소스를 먼저 확인하는 Prompt가 선택됩니다. 직접 명령으로 전환하거나 로컬 사본을 다운로드할 수도 있습니다.
설치 여부를 결정하기 전에 SKILL.md와 SkillsMP에 표시된 보조 파일을 읽어 보세요.
Codex 또는 Claude로 설치 이 Prompt를 복사해 Codex, Claude 또는 다른 어시스턴트에 붙여 넣으면 Skill 페이지를 검토하고 설치를 진행할 수 있습니다.
직접 명령은 검토 Prompt를 거치지 않습니다. 실행하기 전에 소스를 확인하세요.
npx skills add https://github.com/oimiragieo/agent-studio --skill notification-triggers명령은 한 줄로 유지됩니다. 복사하기 전에 가로로 스크롤해 전체 내용을 확인하세요.
로컬 사본을 원하시나요? SkillsMP에서 현재 제공할 수 있는 파일을 다운로드하세요.
SOC 직업 분류 기준
SKILL.md 표시 중
| name | notification-triggers |
| description | Configurable regex-based alert system for detecting patterns in tool calls and session activity. Supports error triggers, content regex matching, token threshold triggers, and pattern detection with configurable actions. |
| version | 1 |
| model | sonnet |
| invoked_by | both |
| user_invocable | true |
| tools | ["Read","Bash","Grep","Write","TaskUpdate"] |
| verified | true |
| lastVerifiedAt | "2026-03-23T23:18:08.049Z" |
| best_practices | ["Follow existing project patterns","Document all outputs clearly","Handle errors gracefully"] |
| error_handling | graceful |
| streaming | supported |
| source | builtin |
| trust_score | 100 |
| provenance_sha | bd3ac619db5020d4 |
Read relevant files and understand requirements
Perform the skill's main function using available tools
Return results and save artifacts if applicable
</execution_process>
<best_practices>
</best_practices>
**Example Commands**:# Invoke this skill
/notification-triggers [arguments]
# Or run the script directly
node .claude/skills/notification-triggers/scripts/main.cjs --help
</usage_example>
Notification triggers are defined in .claude/context/runtime/notification-triggers.json:
{
"version": "1.0",
"triggers": [
{
"id": "unique-trigger-id",
"name": "Human-readable trigger name",
"description": "What this trigger detects",
"type":
Condition Fields:
pattern — JavaScript RegExp (error/regex triggers)minTokens — Token threshold for alert (token_threshold trigger)minOccurrences — Pattern match count before triggering (pattern trigger)scope — Where to evaluate: current session, task, or cumulativeAction Types:
log — Write to .claude/context/runtime/notifications.jsonlalert — Display alert in session outputspawn_agent — Create Task() to handle the notificationinterrupt — Pause execution for user inputStep 1: Load Configuration
.claude/context/runtime/notification-triggers.jsonStep 2: Evaluate Conditions For each enabled trigger:
pattern against tool stdout/stderrminTokensminOccurrencesStep 3: Execute Matched Actions If condition is true:
.claude/context/runtime/notification-log.jsonlStep 4: Record Event
{
"timestamp": "ISO-8601",
"triggerId": "id",
"triggerName": "name",
"matched": true,
"context": {
"sessionId": "current session",
"taskId": "current task",
"tokenCount": number
}
}
Log Action
// Append to notifications.jsonl
fs.appendFileSync(
notificationsPath,
JSON.stringify({
timestamp: new Date().toISOString(),
triggerId: trigger.id,
message: renderTemplate(action.message, context),
severity: 'info|warn|error',
}) + '\n'
);
Alert Action
// Write to stdout with standard format
console.log(`\n🔔 NOTIFICATION [${trigger.name}]\n${renderedMessage}\n`);
Spawn Agent Action
// Create async task to handle notification
Task({
task_id: `notify-${trigger.id}-${Date.now()}`,
subagent_type: action.target,
prompt: `Handle notification: ${trigger.description}\nContext: ${JSON.stringify(context)}`,
});
Interrupt Action
// Pause and ask user
const response = await AskUserQuestion({
question: `${trigger.name}: ${action.message}\n\nContinue? [y/n]`,
});
if (response.toLowerCase() !== 'y') {
process.exit(0);
}
{
"id": "token-threshold-warning",
"name": "High Token Usage Warning",
"description": "Alert when session exceeds 120K tokens",
"type": "token_threshold",
"enabled": true,
"condition": {
"minTokens": 120000,
"scope": "session"
},
"actions": [
{
"type": "log",
"message": "Session token count: ${tokenCount}"
},
{
"type": "alert",
"message": "⚠️ Token budget at 60% — consider compression"
},
{
"type": "spawn_agent",
"target": "context-compressor",
"message": "Proactive context compression triggered"
}
]
}
{
"id": "hook-error-pattern",
"name": "Hook Execution Errors",
"description": "Detect repeated hook failures in session",
"type": "pattern",
"enabled": true,
"condition": {
"pattern": "hook.*error|hook.*failed|HOOK_ERROR",
"minOccurrences": 3,
"scope": "session"
},
"actions": [
{
"type": "log",
"message": "Hook error pattern detected: ${patternMatches} occurrences"
},
{
"type": "alert",
"message": "🚨 Hook system unstable — check logs"
}
]
}
{
"id": "security-violation-detect",
"name": "Security Policy Violation",
"description": "Alert on potential security violations",
"type": "regex",
"enabled": true,
"condition": {
"pattern": "shell:\\s*true|JSON\\.parse\\(|eval\\(|exec\\(",
"scope": "task"
},
"actions": [
{
"type": "alert",
"message": "⚠️ Potential security issue detected — review code"
},
{
"type": "spawn_agent",
"target": "security-architect",
"message": "Review potential security issue in task output"
},
{
"type": "interrupt",
"message": "Continue despite security concern?"
}
]
}
For code discovery, follow this priority order:
pnpm search:code "<query>" (Primary intent-based search)ripgrep (for exact keyword/regex matches)Before starting:
cat .claude/context/memory/learnings.md
cat .claude/context/memory/decisions.md
After completing:
.claude/context/memory/learnings.md.claude/context/memory/issues.md.claude/context/memory/decisions.mdASSUME INTERRUPTION: Your context may reset. If it's not in memory, it didn't happen.