Skill

recall

Search and retrieve memories. Use when preparing plan, before architecture decisions, on errors, during reviews, before dispatching agents, or when context from past sessions is needed.

npx claudepluginhub lklimek/agents --plugin memcan

Tool Access

This skill is limited to using the following tools:

mcp__plugin_memcan_brain__searchmcp__plugin_memcan_brain__search_memoriesmcp__plugin_memcan_brain__search_standardsmcp__plugin_memcan_brain__search_codemcp__plugin_memcan_brain__get_memoriesmcp__plugin_memcan_brain__count_memoriesmcp__plugin_memcan_brain__update_memorymcp__plugin_memcan_brain__delete_memory

Preview

Search and retrieve knowledge from past sessions across all collections.

SKILL.md

Similar Skills

memory

Recalls past work, decisions, error solutions, and project history via a token-efficient 3-layer memory search workflow (search, timeline, details).

aitytech-agentkits-memory

memory-management

Manages AI memories: saves decisions/patterns/facts with categories/tags, searches context, lists/deletes entries, rescans project architecture via MCP tools and /remember /forget.

1 file

ai-memory

memory-recall

1.1k

Retrieves relevant memories from past sessions using memsearch for historical context, decisions, debugging notes, and project knowledge. Activates on relevance or '[memsearch] Memory available' hints.

1 tool

memsearch

Stats

Stars0

Forks0

Last CommitMar 16, 2026

Actions

View Source View Plugin View on GitHub View README

Help us improve

Share bugs, ideas, or general feedback.

Recall

Search and retrieve knowledge from past sessions across all collections.

Procedure

Determine query -- extract the key topic, error message, or concept to search for.

Search -- run search(query=..., project=<repo-name>) to search across all collections (memories, standards, code, todos) in one call. Set project to git remote origin repo name (e.g., memcan not memcan-2). Use collections param to narrow scope when needed (e.g., collections=["standards"]).

List all (optional) -- if search returns few results and broader context is needed, use get_memories(project=...) and/or get_memories() for full listings.

Opportunistic cleanup -- if any returned memories are vague, ephemeral, obsolete, or near-duplicates of better memories, fix them on the spot: update_memory to improve, or delete_memory to remove. Do NOT search deliberately for bad memories — only act on what surfaces during normal recall.

Report -- present relevant results to the conversation. Include memory IDs for reference.

When to Invoke

Session start -- load project context and recent decisions

Before architecture decisions -- check for prior art or known pitfalls

On errors -- search for similar past issues and fixes

During reviews -- recall coding conventions and preferences

Before planning -- check for relevant past experience

MCP Tools

Tool

Params

Use

search

query: str, project?: str, collections?: [str], limit?: int

Search all collections in one query.

search_memories

query: str, project?: str, limit?: int

Memories-only search.

search_standards

query: str, standard_id?: str, standard_type?: str, limit?: int

Search indexed standards.

search_code

query: str, file_path?: str, tech_stack?: str, limit?: int

Search indexed code.

get_memories

project?: str, limit?: int

List memories by scope.

count_memories

project?: str

Count memories.

update_memory

memory_id: str, memory: str

Fix a memory encountered during recall.

delete_memory

memory_id: str

Remove obsolete memory.

Recall

Search and retrieve knowledge from past sessions across all collections.

Procedure

Determine query -- extract the key topic, error message, or concept to search for.
Search -- run search(query=..., project=<repo-name>) to search across all collections (memories, standards, code, todos) in one call. Set project to git remote origin repo name (e.g., memcan not memcan-2). Use collections param to narrow scope when needed (e.g., collections=["standards"]).
List all (optional) -- if search returns few results and broader context is needed, use get_memories(project=...) and/or get_memories() for full listings.
Opportunistic cleanup -- if any returned memories are vague, ephemeral, obsolete, or near-duplicates of better memories, fix them on the spot: update_memory to improve, or delete_memory to remove. Do NOT search deliberately for bad memories — only act on what surfaces during normal recall.
Report -- present relevant results to the conversation. Include memory IDs for reference.

When to Invoke

Session start -- load project context and recent decisions
Before architecture decisions -- check for prior art or known pitfalls
On errors -- search for similar past issues and fixes
During reviews -- recall coding conventions and preferences
Before planning -- check for relevant past experience

MCP Tools

Tool	Params	Use
`search`	`query: str, project?: str, collections?: [str], limit?: int`	Search all collections in one query.
`search_memories`	`query: str, project?: str, limit?: int`	Memories-only search.
`search_standards`	`query: str, standard_id?: str, standard_type?: str, limit?: int`	Search indexed standards.
`search_code`	`query: str, file_path?: str, tech_stack?: str, limit?: int`	Search indexed code.
`get_memories`	`project?: str, limit?: int`	List memories by scope.
`count_memories`	`project?: str`	Count memories.
`update_memory`	`memory_id: str, memory: str`	Fix a memory encountered during recall.
`delete_memory`	`memory_id: str`	Remove obsolete memory.