What happens if a data source is missing?

The skill is designed to fail gracefully; if a CLAUDE.md file doesn't exist or the Memory MCP is unavailable, it skips that layer and continues searching the remaining sources.

How does the skill know when to trigger?

It triggers automatically when Claude detects user frustration regarding repeated errors, or manually when the user invokes the /recall command to refresh the session context.

Does it support non-English keywords?

Yes, the search strategy is designed to test both English and Chinese keywords to ensure comprehensive retrieval from all data sources.

How does it handle messy memory data?

The skill includes a mandatory 'Memory MCP Cleanup Checklist' to identify duplicates, merge similar entities, and delete obsolete nodes after every search.

What files does this skill search?

It specifically queries {repo}/CLAUDE.md for project-specific rules and ~/.claude/CLAUDE.md for global user preferences and identity information.

Session Knowledge Recall

Name: Session Knowledge Recall
Author: cuipengfei

bycuipengfei

•

生産性とワークフロー

Retrieves historical project data, user preferences, and learned patterns to ensure consistent and contextually aware AI assistance.

概要

Session Knowledge Recall is a sophisticated multi-tier retrieval system designed to keep Claude perfectly aligned with your project requirements and personal coding style. By querying Project-level CLAUDE.md files, global User configurations, and long-term Memory MCP nodes, it prevents the AI from repeating mistakes or forgetting previously established preferences. This skill is particularly valuable during long-running projects where maintaining context across multiple sessions is critical for productivity and accuracy.

主な機能

Three-tier retrieval from project rules, user identity, and historical experience nodes
Intelligent trigger detection for user frustration, manual refresh, or general queries
5 GitHub stars
Multi-language keyword matching supporting both English and Chinese search terms
Mandatory Memory MCP maintenance to merge duplicates and keep the knowledge graph clean
Structured summaries that prioritize the most specific and relevant project information

ユースケース

Recalling specific project coding conventions and toolchain configurations automatically
Accessing historical problem-solving patterns to avoid retreading failed implementation paths
Synchronizing personal communication styles and quality standards across all Claude sessions

概要

主な機能

Three-tier retrieval from project rules, user identity, and historical experience nodes
Intelligent trigger detection for user frustration, manual refresh, or general queries
5 GitHub stars
Multi-language keyword matching supporting both English and Chinese search terms
Mandatory Memory MCP maintenance to merge duplicates and keep the knowledge graph clean
Structured summaries that prioritize the most specific and relevant project information

ユースケース

Recalling specific project coding conventions and toolchain configurations automatically
Accessing historical problem-solving patterns to avoid retreading failed implementation paths
Synchronizing personal communication styles and quality standards across all Claude sessions