菜单
ALWAYS use this skill instead of Read when reading Claude Code transcript files (.jsonl in ~/.claude/projects/, ~/.claude/history.jsonl, or paths containing 'transcript'). Token-efficient format that extracts USER/ASSISTANT exchanges and tool summaries, strips metadata.
用 Codex 或 Claude 帮你安装 复制这段 Prompt，粘贴到 Codex、Claude 或其他助手里，让它检查 Skill 页面并帮你完成安装。
基于 SOC 职业分类
| name | read-transcript |
| description | ALWAYS use this skill instead of Read when reading Claude Code transcript files (.jsonl in ~/.claude/projects/, ~/.claude/history.jsonl, or paths containing 'transcript'). Token-efficient format that extracts USER/ASSISTANT exchanges and tool summaries, strips metadata. |
| allowed-tools | Bash |
Token-efficient reader for Claude Code .jsonl transcript files. Strips metadata, extracts conversation flow.
CRITICAL: Use this skill AUTOMATICALLY instead of Read for any transcript files.
File Patterns:
~/.claude/projects/*/<UUID>.jsonl - main session transcripts~/.claude/projects/*/agent-*.jsonl - sub-agent transcripts (at project root)~/.claude/projects/*/<UUID>/subagents/agent-*.jsonl - nested sub-agent transcripts~/.claude/history.jsonl - ⚠️ command history (different format, not supported)*/diffs/*/transcript_cycle.jsonl - transcript cycle files.jsonl when context suggests Claude Code conversationUser Phrases:
Only use Read directly if:
# Basic - outputs to stdout
.claude/skills/read-transcript/scripts/read-transcript /path/to/transcript.jsonl
# With metadata header (dir, branch, timestamps, entry count)
.claude/skills/read-transcript/scripts/read-transcript transcript.jsonl --summary
# Text only (no tool calls)
.claude/skills/read-transcript/scripts/read-transcript transcript.jsonl --no-tools
# Compact format
.claude/skills/read-transcript/scripts/read-transcript transcript.jsonl --compact
# Include sub-agent transcripts inline
.claude/skills/read-transcript/scripts/read-transcript transcript.jsonl --inline-subagents
# Pagination (like Read tool's offset/limit)
.claude/skills/read-transcript/scripts/read-transcript transcript.jsonl --offset 10 --limit 5
| Flag | Description |
|---|---|
--summary | Include metadata header (dir, branch, timestamps) |
--no-tools | Omit tool calls/results (text exchanges only) |
--compact | Denser output format |
--inline-subagents | Recursively inline sub-agent transcripts |
--thinking | Include thinking blocks (usually skip to save tokens) |
--offset N | Skip first N entries (0-based, like Read tool) |
--limit N | Return only N entries (like Read tool) |
Auto-trigger on:
.jsonl file in ~/.claude/projects/Also useful for:
Supports Read-like pagination for large transcripts:
# Entries 10-14 (0-based indexing)
.claude/skills/read-transcript/scripts/read-transcript transcript.jsonl --offset 10 --limit 5 --summary
# Output: "Entries: 5 (showing 10-14 of 745)"
Use --summary with pagination to see the range displayed.
Wrapper around codex CLI for non-interactive runs (prompt via args/file/stdin) with resume support. Outputs session_id for conversation continuation.
Multi-agent debate orchestrator. Use when you want multiple AI perspectives on a question, plan review, or architectural decision. Runs parallel agent panels with a Chair synthesizing responses.
Wrapper around gemini CLI for non-interactive runs (prompt via args/file/stdin) with resume support. Outputs session_id for conversation continuation.
SDK-based Claude agent with full Task tool access. Use for executing handoff prompts or spawning nested sub-agents.
This skill should be used when the user asks to "implement this plan", "generate execution prompts", "super implement", "create handoff from plan", or wants to transform a large plan into self-contained execution-ready prompt artifacts.
Build a phased prompt chain with orchestration file from a monolithic implementation prompt