Rechercher

PythonReactSEOPDF
GitHub Skill

recall

Search the Nori knowledge base for relevant context, solutions, and documentation.

tilework-tech☆ 106 Étoiles du dépôt5 Forks du dépôt

Installation

Suivez les instructions du dépôt pour votre agent. Consultez les fichiers et dépendances avant l’installation.

Voir la source ↗Guides

Instructions

Recall

Searches the shared knowledge base for relevant context or fetches specific articles by ID.

When to Use

Search mode - Search for:

  • Previous solutions and debugging sessions
  • User-provided docs and project context
  • Code patterns and architectural decisions
  • Bug reports and conventions

Fetch mode - Retrieve full article when:

  • You have an article ID from search results
  • You need complete article content (not just snippets)
  • You want to dig deeper into a specific result

Skip recall when:

  • You need on-disk file contents (use Read tool)
  • Information is in recent conversation history
  • Searching for generic programming knowledge

Usage

Recall is a programmatic google search for internal documents. The search mode gives you snippets. The fetch mode lets you see the full data of those snippets. The best way to use this tool is to iterate back and forth between search and fetch mode.

  • Search for something.
  • Fetch the data that makes the most sense based on the snippets.
  • Read the returned articles in full.
  • Decide if there is something else that is worth searching for based on what you learned.
  • Repeat.

If you do not find anything in the first search, simply search again with different terms.

Search Mode
node {{skills_dir}}/recall/script.js --query="Your search query" --limit=10
Fetch Mode
node {{skills_dir}}/recall/script.js --id="article_id"

Examples

Search for articles
node {{skills_dir}}/recall/script.js --query="implementing authentication endpoints" --limit=5
Fetch specific article
node {{skills_dir}}/recall/script.js --id="nori_abc123def456"

Output

Search mode: Returns article snippets (truncated to 500 chars) with metadata and search source breakdown (keyword, fuzzy, vector).

Fetch mode: Returns complete article content without truncation, including full metadata (name, ID, type, repository, timestamps).

Requirements

  • Paid Nori subscription
  • Configured credentials in ~/nori-config.json

Exemples

Consultez les exemples du dépôt source.

Voir la source ↗

Historique source

Date d’indexation
2026-01-20
Mise à jour de l’index
2026-01-28
Branch
main

Les dates concernent ce répertoire. Consultez le dépôt pour l’historique du code.

GitHub ↗

chat Comments (0)

chat_bubble_outline

No comments yet. Be the first to share your thoughts!