Review the standalone Skill operation catalog and its runtime selection boundary.
Direct HTTPS operations
The Skill calls the same authenticated REST surface the SDK uses, so operations available to the Skill are the documented memory operations rather than a separate API.
GET /v1/memories/search?query=current%20task&limit=5
X-API-Key: <your-assigned-api-key>
Agents talk to /v1/skill/operations using a scoped skill_token. These commands cover every supported operation, with the scope each one requires.
node skills/xmemo/scripts/xmemo-skill.mjs remember --content "text" # memory:write
node skills/xmemo/scripts/xmemo-skill.mjs recall --query "query" # memory:read
node skills/xmemo/scripts/xmemo-skill.mjs search --query "query" # memory:read
node skills/xmemo/scripts/xmemo-skill.mjs save-state --key "key" --content "content" # memory:write
node skills/xmemo/scripts/xmemo-skill.mjs restore-state --key "key" # memory:read
node skills/xmemo/scripts/xmemo-skill.mjs todo-add --content "task" # memory:write
node skills/xmemo/scripts/xmemo-skill.mjs todo-list # memory:read
node skills/xmemo/scripts/xmemo-skill.mjs todo-done --id "task_id" # memory:write
node skills/xmemo/scripts/xmemo-skill.mjs expense-add --item "item" --amount 10 # ledger:write
node skills/xmemo/scripts/xmemo-skill.mjs doctor # no scope; bad credentials fail
Machine-readable Skill discovery
An agent can discover the Skill runtime specification and its endpoints by parsing the well-known discovery contracts, rather than being told them.
- Discovery contracts: /.well-known/agent-discovery.json and /.well-known/memory-os.json
- Runtime model: standalone_skill.runtime_model = "standalone_skill"
- No dependencies: mcp_required = false, npm_cli_required = false
- Operations endpoint: /v1/skill/operations
- Credential type: skill_token
curl -s https://xmemo.dev/.well-known/agent-discovery.json