Append structured architectural decisions and insights to memory.jsonl.
npx skills add https://github.com/markmdev/meridian --skill memory-curator
Please help me install this Agent Skill. Skill: memory-curator Source: https://github.com/markmdev/meridian/tree/main/.claude/skills/memory-curator Command: npx skills add https://github.com/markmdev/meridian --skill memory-curator
Claude often struggles with context loss and forgetting past decisions or lessons learned across sessions. This Skill provides a persistent, searchable memory for critical architectural decisions, problems solved, and insights gained, ensuring that valuable knowledge is never lost and preventing the AI from repeating mistakes.
memory.jsonl.Use the memory-curator skill to add a new memory entry. The summary is "Refactored authentication flow to use JWTs for better scalability." Tag it with "architecture,decision". Link it to "TASK-091 src/auth/jwt.ts".
High-intent search queries and answers about installing and using this skill.
Document architectural decisions and lessons learned by persisting them into memory.jsonl using the add_memory_entry.py script. Each entry captures the decision, root cause, solution, and implications with a mem-XXXX ID and UTC timestamp, creating a searchable knowledge base that prevents repeating past pitfalls.
Use structured memory entries to capture insights, problems solved, and patterns discovered. The memory-curator Skill enforces consistent JSON formatting with summary, tags, and file links, ensuring knowledge remains accessible and queryable for future work without context loss.
Yes. Each memory entry includes links to relevant code locations and a UTC timestamp. The memory.jsonl format supports tagging decisions with categories like architecture or performance, making it easy to retrieve and reference related decisions across your codebase history.
Memory entries are tagged with keywords like architecture, decision, or performance and stored in memory.jsonl with consistent JSON structure. Query by tags or links to quickly surface relevant lessons learned and architectural decisions from previous work.
The memory-curator Skill stores insights as JSON objects appended to memory.jsonl with fields for summary, tags, links to code files, a mem-XXXX ID, and UTC timestamp. This structured format makes entries machine-readable and easy to parse for future analysis.
No. The add_memory_entry.py script enforces a strict workflow that handles formatting automatically. You provide the summary, tags, and file links; the script generates the mem-XXXX ID, timestamp, and JSON structure before appending to memory.jsonl.
growvvStore and retrieve persistent memory entries in a JSONL file.
whjelmarSearches persistent memory markdown files for session summaries, decision records, and knowledge base entries.
roberto-melloCapture solved problems as structured JSONL entries in .beads/memory/knowledge.jsonl.
xz1220Maintain a persistent markdown knowledge base with keyword-based wiki page retrieval.