Skip to content
cccskills
FreePublished on GitHub

memory-recall

Search and recall memories from claude-mem

The English summary is being prepared. Showing the original description.

See how to install

Files included (1)

  • SKILL.md4.2 KB

SKILL.md (original)

Review the instructions the agent will receive before you install.

Memory Recall Skill

Search and recall relevant memories from claude-mem using semantic search.

Parameters

NameTypeRequiredDescription
querystringyes*Search query (*not required with --recent)

Options

--recent, -r     Recall most recent memories
--limit, -l      Maximum results to return (default: 5)
--verbose, -v    Show full memory content
--date, -d       Filter by date (YYYY-MM-DD)

Workflow

1. Build semantic query
   ├── Prefix with "my-project"
   ├── Add user query terms
   └── Include date if specified

2. Search claude-mem
   └── chroma_query_documents

3. Format results
   ├── Sort by relevance score
   ├── Show summary for each
   └── Provide memory IDs for detailed access

4. Present to user

Query Tips

effective_queries:
  - "authentication flow" (topic-based)
  - "2025-01-20 oauth" (temporal + topic)
  - "decision jwt" (type + topic)
  - "agent creation workflow" (semantic intent)

avoid:
  - Single generic words like "code" or "task"
  - Very long queries (keep it focused)
  - Special characters or complex syntax

Retrieval Strategy

Recall-Precision Tradeoff

Default bias: recall > precision — it is easier to filter out irrelevant results (false positives) than to recover missed information (false negatives).

Task TypeBiasRecommended --limitRationale
Debugging / InvestigationHigh recall (16:1)10-15Cast wide net, prune later
Decision referenceBalanced5 (default)Moderate breadth with manageable noise
Specific fact lookupHigh precision3Narrow, targeted retrieval

Guidelines

  • Over-retrieve, then filter: When uncertain, request more results and discard irrelevant ones in post-processing
  • Narrow progressively: Start broad, refine query only if results are noisy — avoid starting too narrow
  • Combine temporal + semantic: Add date filters (--date) to semantic queries for better precision without sacrificing recall

Output Format

Basic Search

[sys-memory-keeper:recall authentication]

Searching memories for: "my-project authentication"

Found 3 relevant memories:

[1] mem_abc123 (Score: 0.95)
    Date: 2025-01-20
    Summary: Implemented OAuth flow with Google provider
    Tags: [authentication, oauth, google]

[2] mem_def456 (Score: 0.87)
    Date: 2025-01-18
    Summary: JWT token decision - RS256 algorithm
    Tags: [authentication, decision, jwt]

[3] mem_ghi789 (Score: 0.82)
    Date: 2025-01-15
    Summary: Authentication architecture discussion
    Tags: [authentication, architecture, planning]

Use "sys-memory-keeper:recall --verbose" for full content.

Verbose Output

[sys-memory-keeper:recall authentication --verbose]

Searching memories for: "my-project authentication"

[1] mem_abc123 (Score: 0.95)
    Date: 2025-01-20
    Tags: [authentication, oauth, google]

    Content:
    ## Session Summary
    Date: 2025-01-20

    ### Tasks Completed
    - Implemented OAuth flow with Google provider
    - Added callback handler for OAuth response
    - Created user session management

    ### Decisions Made
    - Use passport.js for OAuth handling
      Rationale: Well-maintained, good documentation

    ---

[2] mem_def456 (Score: 0.87)
    ...

Recent Memories

[sys-memory-keeper:recall --recent]

Fetching recent memories for: my-project

[1] mem_xyz999 (Score: 1.00)
    Date: 2025-01-24
    Summary: Memory system Phase 1 implementation
    Tags: [session, memory, phase1]

[2] mem_xyz888 (Score: 0.98)
    Date: 2025-01-23
    Summary: Agent identification rules update
    Tags: [session, rules, enforcement]

...

No Results

[sys-memory-keeper:recall "nonexistent topic"]

Searching memories for: "my-project nonexistent topic"

No memories found matching your query.

Suggestions:
- Try different keywords
- Use broader search terms
- Check available memories with "sys-memory-keeper:recall --recent"

Related

  • memory-save - Save current context

Reviews

No reviews yet. Tell others what you thought.

Comments

No login needed. Share how it worked for you, ask a question, or note where it did not run.

No comments yet. Be the first to write one.

Comments are published right away. Do not include personal information such as email addresses or phone numbers. Up to 2 URLs.

0 / 1,000 characters

This form is protected by reCAPTCHA, and the Google Privacy Policy and Terms of Service apply.

More skills from this repository

Pre-action boundary checking — validates agent tool calls against declared capabilities and task contracts

The English summary is being prepared. Showing the original description.

baekenough/second-brain15Updated Oct 8, 2026

Auto-detect project context and optimize harness — deactivate unused agents/skills, suggest missing experts, generate project profile

The English summary is being prepared. Showing the original description.

baekenough/second-brain15Updated Oct 8, 2026

Adversarial code review using attacker mindset — trust boundary, attack surface, business logic, and defense evaluation

The English summary is being prepared. Showing the original description.

baekenough/second-brain15Updated Oct 8, 2026

Apache Airflow best practices for DAG authoring, testing, and production deployment

The English summary is being prepared. Showing the original description.

baekenough/second-brain15Updated Oct 8, 2026

Alembic migration patterns for naming conventions, safety checks, expand-contract, env.py configuration, and CI integration

The English summary is being prepared. Showing the original description.

baekenough/second-brain15Updated Oct 8, 2026

Pre-routing ambiguity analysis — scores request clarity and asks clarifying questions when needed (inspired by ouroboros)

The English summary is being prepared. Showing the original description.

baekenough/second-brain15Updated Oct 8, 2026

See all skills by baekenough

Report a problem with this skill