Scan prompt files, CLAUDE.md docs, skill definitions, and other documentation to seed trigger phrases — bootstrapping total-recall from existing project knowledge. Supports --models flag for model-scoped triggers.
From total-recallnpx claudepluginhub brewpirate/zen-flow --plugin total-recallThis skill uses the workspace's default tool permissions.
Provides UI/UX resources: 50+ styles, color palettes, font pairings, guidelines, charts for web/mobile across React, Next.js, Vue, Svelte, Tailwind, React Native, Flutter. Aids planning, building, reviewing interfaces.
Fetches up-to-date documentation from Context7 for libraries and frameworks like React, Next.js, Prisma. Use for setup questions, API references, and code examples.
Guides Payload CMS config (payload.config.ts), collections, fields, hooks, access control, APIs. Debugs validation errors, security, relationships, queries, transactions, hook behavior.
Automatically discover and scan documentation, prompt files, skill definitions, and configuration docs in the project. This seeds the trigger index from existing project knowledge so total-recall is useful immediately without manually scanning every file.
--models <list>: Comma-separated list of target models (e.g., --models sonnet,opus). Default: sonnet--force or --rescan: Re-scan files that already have triggersGlob for documentation and prompt files across the project:
**/CLAUDE.md — project instructions**/.claude/**/*.md — claude config docs**/skills/**/SKILL.md — skill definitions**/agents/*.md — agent definitions**/commands/*.md — command definitions**/docs/**/*.md — documentation directories**/README.md — readmes**/*.prompt.md, **/*-prompt.md — prompt files**/ARCHITECTURE.md, **/CONTRIBUTING.md, **/CHANGELOG.mdExclude: node_modules/, dist/, build/, .git/
If the user provided a specific path, scan only that path.
Read .claude/triggers.json. Skip any files that already have trigger entries for ALL requested models, unless the user passed --force or --rescan. Files with triggers for some but not all requested models will be scanned for the missing models only.
For each batch of unscanned files (up to 20 per batch), spawn the researcher agent (use the Agent tool with subagent_type set to total-recall:researcher):
Generate trigger phrases for these files.
Target models: <comma-separated model list>
Files:
<list absolute file paths, one per line>
Merge results into .claude/triggers.json using the model-scoped format (version 2). Tag seeded entries with "source": "seed":
{
"src/auth/middleware.ts": {
"models": {
"sonnet": {
"phrase": "jwt route authentication guard",
"samples": ["...", "..."],
"convergence": ["jwt", "route", "auth"],
"confidence": 0.85,
"phases": 1
}
},
"crossModelTerms": ["jwt", "route"],
"source": "seed",
"scannedAt": "2026-04-01T10:00:00.000Z"
}
}
After storing all results, invoke the total-recall:index skill to rebuild the master word index.
Seeded N new triggers from M files (models: sonnet, opus):
File | Model | Trigger Phrase | Confidence
--------------------------------- | ------ | --------------------------------- | ----------
CLAUDE.md | sonnet | project rules and conventions | 0.88
CLAUDE.md | opus | project conventions enforcement | 0.95
Skipped K files (already scanned for all target models). Use --rescan to regenerate.
Rebuilt recall index: X words → Y files across Z models.