This one generates a diagnostic markdown report from your GitHub Copilot CLI or Claude Code session transcript. It auto-detects which harness you're running, parses the JSONL, and breaks down turn counts, token usage, skill invocations, build attempts, and stuck patterns like repeated file reads or build loops. The report includes your full session transcript with prompts and file contents, so there's a big privacy warning baked in reminding you to redact before sharing. Useful when you want to understand why a session went sideways, document what happened during a long build, or spot tooling issues worth fixing. The script can target the current session by ID or fall back to the most recent transcript on disk.
npx -y skills add microsoft/win-dev-skills --skill winui-session-report --agent claude-codeInstalls into .claude/skills of the current project.
Select a file.
sickn33/antigravity-awesome-skills
kubesphere/kubesphere
supercent-io/skills-template