Skill Claude CodeCodex
Judge a denkeeper eval run's blinded response pairs over MCP. Use when the user asks to judge an eval run, work the judging queue, calibrate the judge, or decide whether a candidate model is an upgrade.
Instructions file
Instructions for Temikus/denkeeper, covering claude.md, build & development commands, architecture, conventions and testing patterns.
Plugin Claude Code
Auto-runs a critical session post-mortem when Claude Code stops. Ships the analyze-session command, session-analyzer agent, and a Stop hook that triggers the review in-session (no external API calls). Requires Node.js 20 or newer on PATH.
Agent
Critically analyzes a Claude Code session. Investigates the changes made, along with reading a condensed transcript file and evaluates goal achievement, efficiency, code quality, and provides actionable recommendations. Used by the claude-watchdog Stop hook.
Hook
Runs when the agent finishes a response, executing session-analysis.mjs via node. From Temikus/claude-watchdog.
Hook
Runs when you submit a prompt, before the agent sees it, executing hold-input.mjs via node. From Temikus/claude-watchdog.
Hook
Runs when a subagent finishes for session-analyzer tool calls, executing persist-analysis.mjs via node. From Temikus/claude-watchdog.
Skill Claude CodeCodex
Critically analyze the current session.