Skill

transcript

Generate an HTML transcript of a Claude Code or Claude web session using claude-code-transcripts. Use when the user says "generate a transcript," "save this session," "export this conversation," "make a transcript," or wants to convert a session to HTML.

npx claudepluginhub czue/agent-tools --plugin agent-tools

Tool Access

This skill is limited to using the following tools:

Bash(uvx --from*claude-code-transcripts*)AskUserQuestion

Preview

Generate an HTML transcript of a Claude Code or Claude web session using `uvx --from /home/czue/src/lib/claude-code-transcripts claude-code-transcripts`.

SKILL.md

Similar Skills

skill-lookup

159.9k

Searches, retrieves, and installs Agent Skills from prompts.chat registry using MCP tools like search_skills and get_skill. Activates for finding skills, browsing catalogs, or extending Claude.

prompts.chat

prompt-lookup

159.9k

Searches prompts.chat for AI prompt templates by keyword or category, retrieves by ID with variable handling, and improves prompts via AI. Use for discovering or enhancing prompts.

prompts.chat

next-compile

139.2k

Checks Next.js compilation errors using a running Turbopack dev server after code edits. Fixes actionable issues before reporting complete. Replaces `next build`.

1 file

vercel-next-js-2

Stats

Stars0

Forks0

Last CommitMar 4, 2026

Actions

View Source View Plugin View on GitHub View README

Generate Transcript

Generate an HTML transcript of a Claude Code or Claude web session using uvx --from /home/czue/src/lib/claude-code-transcripts claude-code-transcripts.

All commands use the local fork:

uvx --from /home/czue/src/lib/claude-code-transcripts claude-code-transcripts <command>

Arguments

$0: (optional) output mode — one of open, gist, all, or web

Workflow

Determine the output mode:
- If $0 is provided and is a recognized mode (open, gist, all, web), use it.
- Otherwise, default to open mode (Claude Code session opened in browser).
Ask for the output detail level (for single-session modes only):

Use AskUserQuestion to ask which output mode the user wants:
- Full (Recommended) — all content including tool calls, tool results, and thinking
- Compact — hides tool results but shows tool calls and thinking
- Conversation — text only, hides tool calls, tool results, and thinking
This maps to the --output-mode flag: full, compact, or conversation.
For single-session modes (open and gist), select the session:

a. List recent sessions using the list subcommand:
```
uvx --from /home/czue/src/lib/claude-code-transcripts claude-code-transcripts list --json-output --limit 10
```
This returns JSON with path, summary, date, size, and project for each session.

b. Present the sessions to the user using AskUserQuestion with options built from the JSON. Format each option label like: "2026-03-04 15:12 — can you make yourself a skill to generate transcripts..." Use the path field to identify the selected session.

Run the appropriate command based on the mode:

Open in browser (open):

uvx --from /home/czue/src/lib/claude-code-transcripts claude-code-transcripts json --open --output-mode <mode> "<selected_path>"

Gist (gist, the default):

uvx --from /home/czue/src/lib/claude-code-transcripts claude-code-transcripts json --gist --output-mode <mode> "<selected_path>"

All:

uvx --from /home/czue/src/lib/claude-code-transcripts claude-code-transcripts all --open

Converts all local sessions to a browsable archive and opens it.

Web:
```
uvx --from /home/czue/src/lib/claude-code-transcripts claude-code-transcripts web --open
```
Note: this requires an interactive picker and may not work in Claude Code's terminal. Warn the user and try it anyway.

Report results: Tell the user what was generated and where the output went. If using gist mode, share the Gist URL from the command output.

Notes

The json subcommand is used instead of local to bypass the interactive picker (no TTY in Claude Code).
For gist mode, the user needs gh (GitHub CLI) authenticated.
The --open flag auto-opens results in the browser.