memory
/memory
Reads and writes the project's memory (the Obsidian vault and the docs/ folder). Use it when you need earlier context, past decisions, lessons or project conventions, or when there is something new worth remembering for future sessions.
- Author
- alejandro-xoxo ↗
- Repository
- agent-skills ↗
- Licence
- MIT
- Last repository change
- 12.11.2026
Assessment and security
- Files in this package
- 1
- Scripts
- 0
- HTML comments
- 0
- URLs in SKILL.md
- 0
No signals detected. Repository changed on 12.11.2026, without references to retired models or APIs.
MIT, open: use and modify while retaining author attribution.
What it does and when to use it
Reads and writes the project's memory (the Obsidian vault and the docs/ folder). Use it when you need earlier context, past decisions, lessons or project conventions, or when there is something new worth remembering for future sessions.
Original author description
Reads and writes the project's memory (the Obsidian vault and the docs/ folder). Use it when you need earlier context, past decisions, lessons or project conventions, or when there is something new worth remembering for future sessions.
How it works
- To look something up (in this order)
- To write
- What NOT to store
- Decisions
SKILL.md preview
--- name: memory description: Reads and writes the project's memory (the Obsidian vault and the docs/ folder). Use it when you need earlier context, past decisions, lessons or project conventions, or when there is something new worth remembering for future sessions. --- # Project memory > **Requirement:** the project must have the `## Memory` block in `AGENTS.md`. If it does not, the memory is not set up: use the `iniciar-memoria` skill first (or tell the user) and do not go on. **Paths:** read the `## Memory` block of `AGENTS.md`. It holds the vault path and the folders (index, inbox, lessons, decisions, sources). Below they are called `<vault>`, `<index>`, `<inbox>`, `<lessons>`, `<decisions>` and `<sources>`. If the vault is outside the working directory and you cannot read it, ask the user for access before going on. The memory is layered. Enter through the cheapest layer and go deeper only if needed. ## To look something up (in this order) 1. `docs/STATE.md`: where the work stands. 2. `<index>`: follow only the `[[links]]` the task needs. 3. To find something specific: `grep -ril "word" "<vault>" --include=*.md` and open only the notes that come up. 4. `<sources>` is raw material. Never read it whole; search with grep and read the fragment. 5. Before using a stored fact (path, function, figure, status), check it against the source. If it is no longer true, fix the note. ## To write First look for an existing note on the subject (`grep`). If there is one, update it; do not duplicate. | What it is | Where | Format | |---|---|---| | Loose idea or question | `<inbox>` | One line or a paragraph with a date | | How the domain or process works | the vault, in its domain folder | One idea per note | | Lesson learned | `<lessons>` | The vault's `Lesson.md` template, if it exists | | Decision with reasons | `<decisions>` | The vault's `ADR.md` template, numbered `ADR-NNNN Title.md` | | Raw material | `<sources>` | Unedited copy + a line in its index | Writing rules: - Front matter with `tags` and `created`. For decisions, also `status` (proposed, accepted, superseded). - One idea per note. Clear title. File names without `: ? " < > | / \`. - Store the **why** and the **source** with its date, not just the fact. - Link with `[[wikilinks]]` and add the new note to the index it belongs to. - Write in the language of the project and without people's names or personal data unless essential. - Never delete or overwrite existing notes without asking. ## What NOT to store - What is already in the code or in git. - Secrets, keys or passwords. - Personal data. - Ephemeral state of a single session (that goes to `docs/STATE.md`). ## Decisions If the decision changes how something is built, create an ADR note and add 5 lines to `docs/DECISIONS.md` with the link.
Versions
| Repository | Score | Licence |
|---|---|---|
| This page · alejandro-xoxo/agent-skills | 90 | MIT |
| rjmurillo/ai-agents | 100 | MIT |
| rjmurillo/ai-agents | 100 | MIT |
| topoteretes/cognee-integrations | 100 | Apache-2.0 |
| topoteretes/cognee-integrations | 100 | Apache-2.0 |
| topoteretes/cognee-integrations | 100 | Apache-2.0 |
| thebtf/engram | 100 | MIT |
| stxrmik/antigravity-plus | 100 | MIT |
| boshu2/agentops | 100 | Apache-2.0 |
Database snapshot: 13.11.2026 · package version 637de1547652 (SHA-256 digest of the entire package)
Downloads in the last 30 days: 0 · ranking
Report a problem with this skill: harmful content, incorrect descriptions or an author’s removal request.
Similar skills
- 100sugar-memory
Guides an agent to use Sugar's persistent memory layer to store and surface project context across sessions. Use when starting a task in a project that may have prior context (recall relevant memories, get project context); before making a decision that could benefit from past… - 100oh-my-obsidian-recall
Use this skill when the user asks about previous decisions, earlier fixes, past session notes, troubleshooting history, or knowledge that should already exist in the Obsidian vault. - 90daftari-memory
Use the user's Daftari vault as long-term memory. Use when the user refers to earlier decisions, preferences, or project context; asks you to remember, forget, or correct something; or when a conversation reaches a conclusion worth keeping for future chats. - 100rc-memory
Project memory under .rc/memory/ — facts, lessons, golden paths. Use before working and after a correction. Do not use to edit source. - 90agent-memory
Recall durable project decisions, capture explicitly authorized lessons, and review memory history or corrections using the bundled local memory server. Use when continuity across sessions matters; skip standalone tasks that need no prior context. - 95pact-memory
Persistent memory for PACT agents. Save context, goals, lessons learned, decisions, and entities. Semantic search across sessions. Use when: saving session context, recalling past decisions, searching lessons. Triggers: memory, save memory, search memory, lessons learned, remember, recall