transcript-replay — Claude Code skill: what it does, how to install
transcript-replay walks Claude Code session transcripts turn-by-turn from Agent Monitor data, summarizing user prompts, assistant decisions, and tool executions.
MIT ★ 1,198 updated 2026-10-23
/plugin marketplace add hoangsonww/Claude-Code-Agent-Monitor
/plugin install ccam-sessions@claude-code-agent-monitor-pluginsWhat transcript-replay does
The skill pulls dialogue history from the local Agent Monitor API using GET /api/sessions/:id and GET /api/sessions/:id/transcript. It begins by printing a session header: ID, model, workspace basename, status, turn count, thinking blocks count, duration in milliseconds, and total cost in USD rounded to four decimals.
Next, it reconstructs the sequence into an intent-action-result structure. Long text snippets exceeding ~200 characters are trimmed with an ellipsis. A concluding section lists the starting goal, major turning points, failed tool calls, and final status.
Without a running Agent Monitor instance started via npm start in the monitor repository root, the skill cannot retrieve session records. Arguments can target a specific session ID, latest, last, a turn range such as 1-10, or the errors keyword.
What it helps with
- inspect long agent sessions without browsing raw JSON logs
- isolate tool failures and execution retries inside a conversation
- audit token expenditure and duration metrics per session
- reconstruct assistant decision chains from prompt to end state
Example prompts for the agent
- Replay the latest session using transcript-replay
- Check tool errors in session 4a2b9c with transcript-replay errors
- Summarize turns 1-10 of the last session using transcript-replay
How to install transcript-replay
1Plugin marketplace
/plugin marketplace add hoangsonww/Claude-Code-Agent-Monitor/plugin install ccam-sessions@claude-code-agent-monitor-pluginsThis installs the whole ccam-sessions plugin with all its skills
2Skills CLI (Vercel)
npx skills add hoangsonww/Claude-Code-Agent-Monitor --skill transcript-replay -a claude-code3Manually: put the files in place
~/.claude/skills/transcript-replay/SKILL.mdCopy the whole skill folder; for one project use .claude/skills/transcript-replay/
1Skills CLI (Vercel)
npx skills add hoangsonww/Claude-Code-Agent-Monitor --skill transcript-replay -a codex2Manually: put the files in place
.agents/skills/transcript-replay/SKILL.mdFor all projects use ~/.agents/skills/
1Skills CLI (Vercel)
npx skills add hoangsonww/Claude-Code-Agent-Monitor --skill transcript-replay -a cursor2Manually: put the files in place
.cursor/skills/transcript-replay/SKILL.mdCursor also reads .agents/skills/ and .claude/skills/
1GitHub CLI
gh skill install hoangsonww/Claude-Code-Agent-Monitor plugins/ccam-sessions/skills/transcript-replay2Manually: put the files in place
.github/skills/transcript-replay/SKILL.md1Skills CLI (Vercel)
npx skills add hoangsonww/Claude-Code-Agent-Monitor --skill transcript-replay -a gemini-cli2Platform CLI
gemini skills install https://github.com/hoangsonww/Claude-Code-Agent-Monitor --path plugins/ccam-sessions/skills/transcript-replayManually: put the files in place
.devin/skills/transcript-replay/SKILL.mdThe legacy .windsurf/skills/ path still works
Claude.ai
Zip the folder and upload it in Settings → Capabilities → Skills
Step by step, with commands for each platform: How to install skills in Claude Code: five ways
Questions about transcript-replay
How to install transcript-replay in Claude Code?
Register the source repository by running /plugin marketplace add hoangsonww/Claude-Code-Agent-Monitor, then install the package with /plugin install ccam-sessions@claude-code-agent-monitor-plugins.
What does transcript-replay do?
It queries the Agent Monitor API to generate a compact, turn-by-turn recap of user prompts, agent choices, tool results, and session costs.
Why does transcript-replay report an unreachable dashboard?
The skill requires the monitor server running locally. Start it with npm start inside the Agent Monitor repository root.
Source text
SKILL.md · MIT
--- name: transcript-replay description: > Walk a Claude Code session transcript turn-by-turn from Agent Monitor data, summarizing each user, assistant, and tool message in order so a long conversation can be reviewed quickly. Anchors the recap to the session header (model, cost, turn_count). Use when reviewing what was actually said and done in a conversation. --- # Transcript Replay Replay a session transcript one turn at a time with a concise summary of each message. ## Input The user provides: **$ARGUMENTS** - A **session ID** to replay, or - "latest" / "last" for the most recent session. - Optionally a turn range (e.g. `1-10`) or `errors` to focus on tool failures. ## Data Sources | Endpoint | Returns | |----------|---------| | `GET /api/sessions/:id/transcript` | ordered transcript messages: role (user / assistant / tool), content, tool_name and tool result where applicable, timestamps | | `GET /api/sessions/:id` | session header: status, model, cwd, started_at, ended_at, cost, metadata (thinking_blocks, turn_count, total_turn_duration_ms) | ## Report Sections ### 1. Header `GET /api/sessions/:id`. Print id, model, cwd basename, status, turn_count, thinking_blocks, total_turn_duration_ms, and cost in one block. ### 2. Turn-by-turn walk `GET /api/sessions/:id/transcript`. Iterate messages in order. For each turn emit one compact entry: - **user** — the request in one sentence (quote the literal ask only if short). - **assistant** — the decision / action taken, plus which tools it invoked. - **tool** — the tool name and a one-line result (success value or the error text); do not paste large tool payloads. Group an assistant message with the tool calls it triggered so each "turn" reads as intent → action → result. ### 3. Thread highlights After the walk, pull out: the original goal, the key turning points, any tool failures or retries, and how the session ended (resolved / errored / abandoned). ## Output A numbered turn list (`Turn N — <role>: <one-line summary>`), grouped intent → action → result, preceded by the header block and followed by the highlights. Truncate any quoted content past ~200 chars with `…`. Currency as USD to 4 decimal places. Summarize faithfully — never invent message content that is not in the transcript. If the transcript endpoint returns empty, say the session has no stored transcript (it may predate transcript capture or need a reimport) rather than fabricating turns. If the dashboard is unreachable, tell the user to start it with `npm start` from the repo root.
© Son Nguyen · MIT · hoangsonww/Claude-Code-Agent-Monitor