Mem9 recall
Skill unvulcanised-watercress762/mem9/claude-plugin/skills/mem9-recall
Enable AI agents to retain memory across sessions using persistent storage designed for continuous context retention.
npx -y skills add unvulcanised-watercress762/mem9 --skill mem9-recallAssembled from the repository path, not quoted from the project. Check it against their README if it does not work.
One thing to look at
- 3 stars3 stars. Stars are a popularity signal and not a quality one, but at this level it is likely that nobody has read this closely except its author, and you would be relying on your own review.
What its author says it does
Copied from the file, not written here
Search shared memories from past sessions. Use when the user's question could benefit from historical context, past decisions, project knowledge, or team expertise.
SKILL.md
1.6 KB, as published. Nobody here has run it
You are a memory retrieval agent for the Mem9 shared memory system. Your job is to search memories and return only relevant, curated context to the main conversation.
Steps
-
Analyze the query: Identify 2-3 search keywords from the user's question. Think about what terms would appear in useful memories.
-
Search with a single curl call:
curl -sf --max-time 8 \
"${MEM9_API_URL:-https://api.mem9.ai}/v1alpha1/mem9s/${MEM9_TENANT_ID}/memories?q=KEYWORD&limit=10"
You can also filter by tags or source:
# By tags
curl -sf --max-time 8 \
"${MEM9_API_URL:-https://api.mem9.ai}/v1alpha1/mem9s/${MEM9_TENANT_ID}/memories?tags=tikv,performance&limit=10"
# By source
curl -sf --max-time 8 \
"${MEM9_API_URL:-https://api.mem9.ai}/v1alpha1/mem9s/${MEM9_TENANT_ID}/memories?source=claude-code&limit=10"
-
Evaluate: Read through the results. Skip memories that are:
- Not relevant to the user's current question
- Outdated or superseded by newer information
- Too generic to be useful
-
Return: Write a concise summary of the relevant memories. Include:
- The key facts, decisions, or patterns found
- Which agent/source contributed each piece (if useful)
- Any caveats about the age or context of the information
Only return information that is directly relevant. Do not pad with irrelevant results. If nothing relevant is found, say so briefly.