Who it is forA person whose agent-led project has accumulated accepted and rejected ideas until the direction is unclear.
The problemAccepted and rejected ideas pile up, so it is hard to tell which rules must hold and which details can still change.
The outcomeThe conversation is restated with governing principles on one side and adjustable implementation choices on the other, for the user to confirm.
What you need first
- An Agent host that supports the Agent Skills specification, such as one the cross-host skills CLI can install into
- Node.js and npm, for the one-line install and the npm scripts
- Access to the relevant conversation history, since the Skill reports gaps instead of rebuilding missing history
Steps
- Run
npx skills add jinhanbuilds/hoist-the-elephant --skill hoist-the-elephant -y - Reopen the Agent host so the Skill is loaded
- Invoke
$hoist-the-elephantat the point where the thread feels stuck - Mark each restated item as a principle or an adjustable choice
- Confirm the split so the next revision uses only the agreed principles
Prompt for your agent
Use the hoist-the-elephant Skill from jinhanbuilds/hoist-the-elephant on the thread about <project>. Output a list of governing principles and a separate list of adjustable implementation choices, then ask me to confirm. Verify that no choice is moved into the principles without my confirmation.
Caveat. The split is only as complete as the conversation history the agent can read, and the Skill asks you to fix any gaps it finds.
Use this in your AI agent
Ask your agent
Using the Deduped MCP, get the setup steps for jinhanbuilds/hoist-the-elephant and "Split fixed principles from adjustable choices in a stalled thread"
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