Skip to content

README.md

Latest commit

 

History

History
49 lines (33 loc) · 2.13 KB

README.md

File metadata and controls

49 lines (33 loc) · 2.13 KB
Raw

vardoger — OpenClaw Plugin

An OpenClaw skill that analyzes your conversation history and generates personalized instructions.

Prerequisites

Install

The direct setup path tracks the current PyPI release and is recommended:

pipx install vardoger
vardoger setup openclaw

This installs the vardoger analysis skill to ~/.openclaw/skills/vardoger/. OpenClaw discovers it automatically on the next session.

Vardoger is also listed on ClawHub as vardoger-analyze. Version 0.3.2 is live there as the latest tag and passed ClawHub's current moderation scan. Install the CLI with pipx in either case because the skill invokes it at runtime.

Maintainers publish future versions with the current CLI shape:

clawhub skill publish plugins/openclaw/skills/analyze \
  --slug vardoger-analyze \
  --name "vardoger — Analyze History" \
  --version X.Y.Z \
  --tags latest \
  --changelog "Describe the release"

Usage

Once loaded, ask OpenClaw to "analyze my conversation history" or "run the vardoger skill."

Where the personalization lands

  • Default — user-global scope: writes to ~/.openclaw/skills/vardoger-personalization/SKILL.md, which OpenClaw auto-discovers on every session.
  • Opt-in — project scope: pass project_path="<workspace root>" (and scope=project) to land <project>/skills/vardoger-personalization/SKILL.md. vardoger refuses to write project-scoped skills into a directory that doesn't look like a real project (it requires .git, a language manifest, AGENTS.md, or an existing .cursor/ in the path or one of its ancestors). Without that check, an MCP server launched from $HOME would silently drop skills into a location OpenClaw would never read. Supply a real workspace root or drop the project_path argument to write user-globally.

Uninstall

Remove the ~/.openclaw/skills/vardoger/ directory.