Who it is forA small team lead who wants the release process learned once to be reused by a dedicated release agent.
The problemRelease steps live in one person's head or old chat logs, so each release starts from scratch.
The outcomeA versioned Skill with execution steps and validation rules is shared with the team and bound to the Release agent.
What you need first
- Git and a bash-compatible shell
- Two sets of LLM API parameters for the memory group and the proxy group, filled into .env
- A local machine that can run the three services (memory-core, memory-hub, proxy)
- An agent client that accepts a custom base URL, such as Claude Code or Codex
Steps
- Start the stack with
./start-all.shfrom deploy/global-images and open http://localhost:8125 - Import the past conversation session where the release was completed so Skills are extracted
- Review the extracted Release Checklist Skill and check its execution steps and validation rules
- Share the Skill with the team and bind it to the Release agent in Agent Loadout
- Ask the Release agent to run the checklist on a test branch and confirm each step is reported
Prompt for your agent
Using TencentDB-Agent-Memory, extract a Release Checklist Skill from the session <session file> and bind it to the agent <release agent name> through the panel at http://localhost:8125. Expected output: the Skill's version, steps, and validation rules, then a checklist run by the agent. Verify the Skill status is shared with the team before it is bound.
Caveat. Skills are private by default, so the Skill must be explicitly shared with the team before the Release agent can use it.
Use this in your AI agent
Ask your agent
Using the Deduped MCP, get the setup steps for TencentCloud/TencentDB-Agent-Memory and "Save a release checklist as a Skill and assign it to a release agent"
Add Deduped to Claude Code once
claude mcp add --transport http deduped https://dedupedai.com/mcp
Every Monday: the 5 hottest repos and exactly how to use them. Free, one email a week, unsubscribe anytime.
Get them freeShare: X LinkedIn Hacker News Email Copy link
Third-party repos we don't own. AI-written use cases can be wrong. Disclaimer