Agent Plugins Marketplace
← All plugins

agent-transcripts

v0.3.0

Export Claude Code, pi and dsh transcripts into one on-device cache and index them in a derived sqlite database.

Claude CodeAgent Plugins6 Skills

By David AlladaLicense: MIT0 GitHub starsUpdated last week

Directory evidence

Runtimes
Claude Code and Agent Plugins
Parsed components
6 skill or MCP entries
Source updated
Sep 15, 2026
Manifest status
Canonical path parsed

The directory validates manifest shape and source location. It does not execute the plugin or provide a security endorsement. Review the indexing methodology →

Install agent-transcripts for Claude Code

Installs for the current user
claude plugin marketplace add IchenDEV/agent-plugin-mkt
claude plugin marketplace update agent-plugin-marketplace
claude plugin install agent-transcripts@agent-plugin-marketplace

Paste and run these commands in a terminal with Claude Code. They add and refresh the PluginsMP catalog, then install this plugin.

The installer fetches third-party code from the source repository shown on this page. This directory validates manifest structure and source location, but does not perform a security audit; review the manifest, components, and source before installing.

Get the source manually
git clone https://github.com/allada-homelab/agent-harness-marketplace

Clone the source repository, then follow its setup instructions to add the plugin to a compatible client. The plugin root is modules/agent-transcripts/.

Plugin files

modules/agent-transcripts/
├── .claude-plugin/plugin.json
├── plugin.json
├── skills/transcript-export-claude/SKILL.md
├── skills/transcript-export-dsh/SKILL.md
├── skills/transcript-export-pi/SKILL.md
├── skills/transcript-ingest/SKILL.md
├── skills/transcript-query/SKILL.md
└── skills/transcript-sweep/SKILL.md

Included Skills6

transcript-export-claudeskills/transcript-export-claude/SKILL.md

Export Claude Code transcript JSONL from the host config directory and dev-container docker volumes into the on-device agent-transcripts cache. Use when the user wants to "export claude transcripts" or "collect claude chat histories". Read-only, incremental, never writes into a git repo.

transcript-export-dshskills/transcript-export-dsh/SKILL.md

Export dsh (DeepSeek harness) zstd session logs from the host DSH_HOME and dev-container docker volumes into the on-device agent-transcripts cache. Use when the user wants to "export dsh transcripts" or "collect deepseek sessions". Read-only, incremental, never writes into a git repo.

transcript-export-piskills/transcript-export-pi/SKILL.md

Export pi coding-agent session JSONL from its settings and default session directories into the on-device agent-transcripts cache. Use when the user wants to "export pi transcripts" or "collect pi sessions". Read-only, incremental, never writes into a git repo.

transcript-ingestskills/transcript-ingest/SKILL.md

Build or refresh the derived sqlite index over exported Claude Code, pi and dsh transcripts in the agent-transcripts cache. Use when the user wants to "index transcripts", "ingest transcripts into sqlite", or "refresh the transcript index". The database is derived and safe to delete; raw files stay canonical. Querying it is the transcript-query skill.

transcript-queryskills/transcript-query/SKILL.md

Query the agent-transcripts sqlite index to measure tool use, sessions, projects and activity across Claude Code, pi and dsh without loading transcript prose into context. Use when the user wants to "analyze transcripts", "query transcripts", "how often did I use X", "find sessions where …", or any question answered from chat history. Aggregates first, bounded excerpts only where judgment is required.

transcript-sweepskills/transcript-sweep/SKILL.md

Sweep the agent-transcripts sqlite index for sessions worth a closer look and write one JSONL report of flagged tool errors, retry loops, duplicate calls, shell commands a dedicated tool covers, and oversized results or sessions. Use when the user wants to "sweep transcripts", "find problem sessions", "where did tools fail", or a first pass before analyzing agent behavior. Deterministic — a script does the counting, and the model only reports the summary and the report path.

Plugin manifests2

modules/agent-transcripts/.claude-plugin/plugin.json
{
  "name": "agent-transcripts",
  "version": "0.3.0",
  "description": "Export Claude Code, pi and dsh transcripts into one on-device cache and index them in a derived sqlite database.",
  "author": {
    "name": "David Allada",
    "email": "[email protected]"
  },
  "license": "MIT"
}
modules/agent-transcripts/plugin.json
{
  "$schema": "https://agent-plugins.org/schemas/1.0.0/plugin.schema.json",
  "name": "agent-transcripts",
  "version": "0.3.0",
  "description": "Export Claude Code, pi and dsh transcripts into one on-device cache and index them in a derived sqlite database.",
  "author": {
    "name": "David Allada",
    "email": "[email protected]"
  },
  "homepage": "https://github.com/allada-homelab/agent-harness-marketplace/tree/main/modules/agent-transcripts",
  "repository": "https://github.com/allada-homelab/agent-harness-marketplace",
  "license": "MIT",
  "keywords": [
    "agent-skills",
    "claude-code",
    "pi",
    "dsh"
  ]
}

For maintainers

If you maintain this plugin, link to this source-backed listing from your README so users can review its manifest and indexed components.

[agent-transcripts on Agent Plugins Marketplace](https://pluginsmp.com/plugins/agent-transcripts)