← ClaudeAtlas

memory-schemalisted

Your memory server's exact call shape and rules. Auto-loads before any memory WRITE (create/update/link a memory) so the model uses the correct tool shape, required fields, enums, char caps, and linking API — and stops inventing fields. Also load it whenever a memory tool errors with a validation failure before retrying.
konstantinos-malavazos/claude-code-playbook · ★ 2 · AI & Automation · score 66
Install: claude install-skill konstantinos-malavazos/claude-code-playbook
# Memory schema (fill in for YOUR memory server) > Replace every `<PLACEHOLDER>` with your actual memory MCP's tool names and constraints. > The point of this skill is to make memory writes correct on the first try. > > **Running Forgetful?** [`forgetful.SKILL.md`](forgetful.SKILL.md) next to this file is > this skill already filled in for it — call shape, caps, enums, the asymmetric linking > table and the tools that don't exist. `mv forgetful.SKILL.md SKILL.md` and delete this > placeholder. On any other server, read it as a worked example of how much detail earns > its place here, then write your own. ## Call shape - Create: `<create-memory-tool>` with fields `<title>`, `<content>`, `<tags>`, `<...>`. - Update: `<update-memory-tool>` ... - Link: `<link-tool>` — note the argument order/direction if it's asymmetric. - Obsolete: `<mark-obsolete-tool>` with a `<superseded_by>` pointer. ## Required fields & caps - `<field>` — required; max `<N>` chars. - `<field>` — enum, one of `<...>`. Do NOT invent values. ## Tagging rules - Tag **by functionality** (topic / component / symbol), **never by ticket id**. (Put the ticket id in the content for provenance if you want it.) - Include your scope/project tag so `/garden-memory` can find untagged strays. - Respect the tag cap (`<N>` tags max) and the locked vocabulary if you have one. ## Content rules - One atomic fact/conclusion per memory. State **what** and **why**. - Link related memories so one query returns the whole chai