npx claudepluginhub n33kos/claude-plugins --plugin imprintThis skill uses the workspace's default tool permissions.
Guides browser automation with Playwright, Puppeteer, Selenium for e2e testing and scraping. Teaches reliable selectors, auto-waits, isolation to fix flaky tests.
Provides checklists to review code for functionality, quality, security, performance, tests, and maintainability. Use for PRs, audits, team standards, and developer training.
Enforces A/B test setup with gates for hypothesis locking, metrics definition, sample size calculation, assumptions checks, and execution readiness before implementation.
Scan your Claude session history, extract behavioral patterns, and generate a profile at ~/.claude/.imprint.
Before running anything, walk the user through the following configuration questions. Present them one at a time as multiple-choice prompts.
How many sessions should I include in training?
- All sessions — most comprehensive profile, recommended for first run or full retrain
- Last 200 sessions — solid coverage, faster extraction
- Last 50 sessions — quick snapshot of recent behavior
- Custom number — you specify
Default: 1 (All sessions)
Map the choice to --max-sessions: all = 0, otherwise the number they pick.
Which model should I use for synthesis?
- Opus — highest quality, slowest, highest token usage. Best if you want the most thorough and nuanced profile.
- Sonnet (recommended) — strong quality, moderate speed and cost. Best balance for most users.
- Haiku — fastest and cheapest, but produces noticeably shorter and shallower profiles. Fine for a quick draft or if you're on a budget.
Weaker models will produce shorter, less nuanced profiles but will complete faster with lower token usage.
Default: 2 (Sonnet)
Map the choice to --model: opus, sonnet, or haiku.
How long should each synthesis step be allowed to run before timing out?
- 10 minutes — good for most users with <500 sessions
- 20 minutes (recommended) — safe default for large session histories
- 30 minutes — for very large histories or slower connections
- Custom (seconds) — you specify
Each synthesis pass and the final merge step get their own timeout. If you have thousands of sessions or are using Opus, choose a longer timeout.
Default: 2 (20 minutes)
Map the choice to --timeout: 600, 1200, or 1800 seconds, or custom.
Once the user has answered all three questions (or accepted defaults), confirm their choices and run the pipeline.
cd "$PLUGIN_DIR/../.." && python3 scripts/extract-sessions.py --output /tmp/imprint-messages.json --max-sessions <N>
cd "$PLUGIN_DIR/../.." && python3 scripts/synthesize-profile.py /tmp/imprint-messages.json --model <MODEL> --timeout <SECONDS>
After training completes:
~/.claude/.imprint and present a summary of what was learnedrm -f /tmp/imprint-messages.jsonIf training fails at any step, report the error and suggest troubleshooting steps.