2.2 KiB
Memory Gateway Agent Skill Implementation Plan
For agentic workers: REQUIRED SUB-SKILL: Use superpowers:subagent-driven-development (recommended) or superpowers:executing-plans to implement this plan task-by-task. Steps use checkbox (
- [ ]) syntax for tracking.
Goal: Create a reusable AI-agent skill that safely operates the Memory Gateway API through a deterministic Python CLI.
Architecture: Keep procedural guidance in SKILL.md, detailed endpoint schemas in references/api.md, and all HTTP/multipart behavior in one standard-library CLI. Read credentials from environment variables or explicit flags and never persist secrets in the skill.
Tech Stack: Agent Skills format, Python 3 standard library, pytest, Memory Gateway HTTP API.
Task 1: Scaffold the skill
Files:
-
Create:
skill/memory-gateway-agent/SKILL.md -
Create:
skill/memory-gateway-agent/agents/openai.yaml -
Create:
skill/memory-gateway-agent/scripts/memory_gateway.py -
Create:
skill/memory-gateway-agent/references/api.md -
Initialize the standard skill structure with
init_skill.py. -
Remove generated placeholders and keep only required resources.
Task 2: Implement and test the CLI
Files:
-
Create:
tests/test_memory_gateway_skill.py -
Modify:
skill/memory-gateway-agent/scripts/memory_gateway.py -
Write failing tests for environment credentials, JSON requests, multipart uploads, and HTTP errors.
-
Run the focused tests and confirm they fail for missing implementation.
-
Implement the standard-library CLI with commands for health, users, resources, search, add/flush, override, and delete.
-
Run the focused tests and confirm they pass.
Task 3: Author and validate the skill
Files:
-
Modify:
skill/memory-gateway-agent/SKILL.md -
Modify:
skill/memory-gateway-agent/references/api.md -
Modify:
skill/memory-gateway-agent/agents/openai.yaml -
Document the agent workflow, authentication rules, ownership checks, and safe handling of secrets.
-
Document endpoint parameters and CLI examples in the API reference.
-
Generate UI metadata with the official skill-creator script.
-
Run
quick_validate.py, CLI--help, focused tests, and the full project test suite.