# Remember Project Context Across Sessions

Important project details are easy to lose between AI sessions. This skill gives Claude, Codex, and Claude Code a structured local memory workflow.

## Install

```bash
npx skillstore add yamadashy/agent-memory
```

## Metadata

- Status: approved
- Slug: yamadashy-agent-memory
- Skillstore revision: r1
- Version status: missing
- Tree hash: 0579a70cc1a607c0a80213560ccc4b393af1e9053bc9da8fe57bad26f3cbd41e
- Author: yamadashy
- GitHub username: yamadashy
- License: MIT
- Repository: https://github.com/yamadashy/repomix/tree/main/.claude/skills/agent-memory
- Ref: 64ca8af0f54a325752f08bd54e52151061ea659a
- Supported tools: Claude, Codex, Claude Code
- Audit status: complete
- Agent install advisory: allowed
- Manual install advisory: allowed
- Artifact signature: available
- Audit attestation: unavailable
- Human verification: not\_verified
- Risk factors: external\_commands, filesystem
- Quality score: 78
- Quality tier: bronze
- Public page: https://skillstore.pages.dev/skills/yamadashy-agent-memory
- Manifest: https://skillstore.pages.dev/api/skills/yamadashy-agent-memory/manifest

## Capabilities

- Creates Markdown memory files with required summary frontmatter.
- Organizes memories into category folders using kebab-case names.
- Searches memory summaries, tags, and full text with ripgrep.
- Updates, deletes, consolidates, and reorganizes outdated memories.
- Uses a gitignored memories directory for local agent notes.

## Use Cases

- Preserve Debugging Knowledge: Save root causes, fixes, and related files after difficult investigations so later sessions can recover context quickly.
- Track Architecture Decisions: Record decisions, rationale, and follow-up notes in a searchable project memory folder.
- Organize Research Notes: Group findings from repeated research tasks into clear categories with searchable summaries and tags.

## Prompt Templates

### Save A Simple Memory

```
Remember this project fact: [fact]. Save it with a clear summary and the right category.
```

### Recall Related Notes

```
Check your memories for anything related to [topic]. Summarize the most relevant notes before we continue.
```

### Update Existing Context

```
Update the saved memory about [topic] with this new information: [details]. Keep the summary current.
```

### Consolidate A Memory Area

```
Review memories about [area]. Merge duplicates, remove outdated notes, and keep a concise searchable structure.
```

## Limitations

- It stores memories as local files, not in a shared remote database.
- It relies on the agent to choose useful and safe content to save.
- It does not provide encryption, access control, or retention policies.
- It requires shell and filesystem access for the documented workflow.

## Best Practices

- Save only durable context that will help future work.
- Write summaries that clearly state whether the full memory is worth reading.
- Review memories periodically to remove outdated or sensitive content.

## Anti Patterns

- Do not store secrets, credentials, tokens, or private customer data.
- Do not save every conversation detail without clear future value.
- Do not delete or reorganize memories without checking the exact target files.

## Security Audit

- Audited at: 2026-07-21T08:30:58.01\+00:00
- Summary: All 19 static findings are false positives. The reported shell-execution findings identify Markdown code fences and inline code, not executed commands. The documented commands operate only on the skill's local memory directory, and no prompt injection, credential access, network activity, or data-exfiltration intent was found.

## Stats

- Views: 321
- Downloads: 98
- Favorites: 2
- Popularity score: 0