claudeers.
// MCP Servers

coding.sh

Open-source, local-first observability and evals for AI agents: see what Claude Code, Codex, Cursor and others did and what it cost.

Install with your AI

Paste into Claude Code, Cursor, or any agent — it reads the repo and wires the tool into your project.

Install and set up coding.sh (claude-plugin project) into my current project.
Found on https://claudeers.com/codingsh
Repo: https://github.com/jackulau/coding.sh
Homepage/docs: https://coding.sh
Detected install method: claude-plugin → /plugin install coding-sh@jackulau/coding.sh
Category: mcp-servers. Platforms: cli, api, web, mobile.
Read the repo's README for exact setup and env vars, then install it and wire it into my project.

Claudeers Health Verdict:
unknown; community-verified: false. Confirm the source before running anything.
// or install directly (claude-plugin)

⚠ Unverified / not recently updated — review before pasting a run-this config.

/plugin marketplace add jackulau/coding.sh
/plugin install coding-sh@jackulau/coding.sh
// or clone
git clone https://github.com/jackulau/coding.sh

// compatibility

Platformscli, api, web, mobile
Operating systems—
AI compatibilityclaude
LicenseMIT
Pricingopen-source
LanguageTypeScript

Get your FREE $2.50 API credits to access TickAtlas financial data ↗

coding.sh: The Observability Layer for Agents. Agents, subagents and gateways, traced on your machine. Below the headline, the top of one session in the coding.sh viewer: a Claude Code session that raised the OTLP receiver body limit to 8 MB, with its model, duration, estimated cost, a '31 passed' test badge and its first file edits.

coding.sh records what your AI agents do: every prompt, model call, tool call, file change and what it cost.
You read each run in a local viewer and turn the ones that went wrong into evals.
Your sessions stay on your machine.

Docs · Quickstart · Supported agents · What you get · Changelog · Discussions · Report a bug

npx coding.sh                               # try it: finds your agents, asks once, imports past sessions
curl -fsSL https://coding.sh/install | sh   # or install it: same setup, then opens the viewer

A terminal recording. Typing 'npx coding.sh' prints a plan: hooks and the MCP server for Claude Code and Codex CLI, T3 Code imported from its store with no hooks, and the coding.sh hook and config. It asks once, 'Set up coding.sh for 2 agents (4 config files, backed up first) and import past sessions? (Y/n)', the answer is y, and it reports 'Set up Claude Code and Codex CLI (hooks + MCP server)', that recording starts with the next session, and 'Imported 9 past sessions from Claude Code and Codex CLI in 0.2 s'. Then 'npx coding.sh usage --by model' prints a table of sessions, turns, tokens, output tokens, tokens per turn, cache read rate and estimated cost for claude-opus-5-5, claude-sonnet-5 and gpt-6.1-sol. The sessions come from .github/assets/sample-sessions.mjs.
A first run of npx coding.sh 0.1.6 in a fresh home.

Run npx coding.sh again to open the viewer on your sessions.

What it costs: nothing. MIT licensed, runs on your machine, no account.

The coding.sh viewer on one Claude Code session: the prompt as the title, agent, model, start time, duration, tokens with the cache rate, the estimated cost, latency percentiles, a '14 passed' test badge, Pass and Fail outcome buttons and a copyable 'coding.sh eval new' command. Below, the 16 steps with each model call's estimated cost and a timing bar, and the selected file change to usage.ts shown as a one-line diff.
One Claude Code session in the local viewer (coding.sh ui), with the diff it made and the tests it ran.

Each session goes into a SQLite index on your machine, with its latency and tokens. You browse it in a local viewer, query it from the CLI or from your agents over MCP, turn failures into evals, and compare changes with paired statistics. No session data leaves your machine unless you add a remote sink.

No runtime dependencies. Node 22.16 or newer.

Quickstart

  1. Run npx coding.sh, or the install command above. It runs coding.sh init: it finds your agents, shows one plan and asks once.
  2. Start a new agent session. Agents read their hooks when a session starts, so one that was already open may not be recorded.
  3. Run coding.sh to open the viewer. New sessions show up on their own, with no ingest step, and your agents can read them over MCP.

init points each agent's hooks at coding.sh, registers the MCP server where the agent supports it, and backs up every config file it edits. It also imports the sessions your agents already keep on this machine, so the viewer opens on your own history.

CommandWhat it does
coding.sh init --dry-runPrints the plan without writing anything. --diff adds the edits.
coding.sh init --no-importSets up capture and skips the past-session import.
coding.sh ingest --backfillImports past sessions again, for example after adding an agent.
coding.sh uninstallRemoves the hooks and MCP entries again.
npm install, the npx cache and other ways to install

You can also install with npm: npm install -g coding.sh. If you only ran npx coding.sh, the hooks keep recording, but the MCP server runs the copy in the npx cache, so install it before you clear that cache. Installing from source, updates and uninstall: docs/guides/install.md.

What you get

AreaWhat it doesStart with
RecordCaptures each session, including subagents, from hooks and transcripts. Wraps any other agent. Takes OpenTelemetry from frameworks and apps.coding.sh init, coding.sh run -- <cmd>, coding.sh collector
MeasureTurns, tool calls, model calls, tokens, latency and outcome for every session. Flags sessions that edited code and never ran a check.coding.sh sessions, coding.sh usage, coding.sh improve
EvaluateTurns a failed session into a replayable eval. Runs experiments with code graders or calibrated LLM judges. A/B tests prompts, harness files and models with paired statistics.coding.sh eval new, coding.sh exp compare
BenchmarkDraws new cases every day, runs every contender on the same frozen set, and writes a static report.coding.sh bench
ProductionOne server for ingest, the viewer and the API, backed by SQLite, Postgres or ClickHouse. SDKs for API apps.coding.sh serve

The coding.sh session list: 13 sessions from Claude Code, Codex and a bot that labels coding.sh issues and sends OpenTelemetry traces, with agent, model, tool count, cost (~ where it is estimated from list prices), duration and start time. Two rows show a green 'passed' test badge and one has a failed tool call. Above the list, a histogram of sessions over time marks one flagged session in amber; filters on the left narrow by agent, model, repository, flag and tool.

Read your sessions

coding.sh sessions             # recent sessions
coding.sh show latest          # timeline with turns, tools, subagents and latency
coding.sh search "rate limit"  # full-text search over prompts, replies and tool calls
coding.sh usage --by model     # tokens, turns and cache hits
coding.sh stats --by agent     # sessions, failures, cost and latency per agent, with a daily trend
coding.sh diff-calls latest#-1 # what changed since the previous model call, and why the cache missed
coding.sh ui                   # the local viewer

Every command takes --json. coding.sh mcp gives your agents the same data as MCP tools. More in docs/guides/explore.md.

Record: wrappers, OpenTelemetry and the model API proxy
  • coding.sh run -- <cmd> wraps any other agent and snapshots the repo before and after.
  • coding.sh collector takes OpenTelemetry from agent frameworks and apps.
  • coding.sh init --proxy routes Claude Code and Codex through a local API proxy that logs each model call's system prompt, tools, messages and cache use. More in docs/how-to/capture-with-proxy.md.
Evaluate: evals, experiments and paired statistics

Turn a failed session into a replayable eval with coding.sh eval new, and check that its verifier is fair with coding.sh eval validate. Run experiments over versioned datasets with code graders or calibrated LLM judges. coding.sh exp compare and coding.sh compare A/B prompts, harness files and models with paired statistics. coding.sh exp replay resends one logged model call with another model or system prompt. Tool evals and multi-agent workflow runs are covered too. More in docs/guides/evals.md.

Benchmark: daily cases and head-to-head reports

coding.sh bench draws new cases every day from a script, your failed sessions, GitHub issues or a sitemap. Every contender gets the same frozen set in random order, and bench report writes a static page with the trend, today's head to head and where each one fails. More in docs/guides/benchmarks.md.

Production: server, SDKs and LLM gateways

coding.sh serve runs ingest, the viewer and the API on one port, backed by SQLite, Postgres or ClickHouse. Your API apps send traces with the TypeScript SDK (@codingdotsh/sdk), the Python SDK (coding-sh) or plain OpenTelemetry. coding.sh gateway links calls made through LLM gateways (Vercel AI Gateway, OpenRouter, LiteLLM, Cloudflare, Portkey, Helicone) to their billed cost. The TypeScript SDK and the store plugins are on npm, and the Python SDK is on PyPI. Start with docs/production/README.md.

Supported agents

An animated diagram from the coding.sh front page. On the left, three columns of agent logo tiles slide past one another. A line joins them to a cube in the middle, which is coding.sh. From the cube, a green line runs to three columns of tiles on the right, labeled sessions, cost, evals, experiments, benchmarks, gateways and sdk.
Every agent, one record: the diagram from the coding.sh front page.

AgentHow it recordsStatus
Claude Codehooks, transcript, subagentsVerified on real sessions
Codex CLIhooks, rollout files, subagentsVerified on real sessions
Antigravity CLI (formerly Gemini CLI)hooks, transcriptVerified on real sessions
OpenCodeplugin, session databaseVerified on a real install
Cursor CLIhooks, transcriptNot run live
GitHub Copilot CLIhooks, session stateVerified on a real install
Factory Droidhooks, transcriptVerified on a real install
Gooseplugin hooks, session databaseVerified on a real install
Piextension, transcriptVerified on a real install
AmppluginNot run live
T3 Codeits local storeVerified on a real install
Aidercoding.sh run -- aider …Verified on a real install
Any other commandcoding.sh run -- <cmd>Tested
Your own agent or app (Vercel AI SDK, Claude Agent SDK, OpenAI Agents SDK, LangGraph and more)OpenTelemetry into coding.sh collector, or the TypeScript and Python SDKsTested
What the statuses mean
  • Verified on real sessions. Checked against sessions captured from a live install of the agent during development, in addition to the CI tests.
  • Verified on a real install. The agent's current release was installed and run through a session against a local stub model, its hooks fired and the recorded session was checked, and the captured payloads are now CI fixtures. For OpenCode, Pi, Goose, Factory Droid, GitHub Copilot CLI and Aider, CI repeats that run on Linux and Windows on every change.
  • Not run live. CI checks the adapter against the agent's documented formats, but nobody has run it against a live install yet. Cursor CLI and Amp are in that state because both need an account login.
  • Tested. Covered by CI tests. There is no third-party agent to check against.

If you use an agent marked "Not run live", a short agent support report saying whether it works moves it to "Verified". Per-agent details, including which agents get the MCP server, are in docs/guides/agents.md. T3 Code and Pi each have a section on Record coding agents.

Privacy

With no sinks configured, no session data leaves your machine. The only default network request is a daily update check, which CODING_SH_NO_UPDATE_CHECK=1 turns off. Secrets are redacted at ingest, before anything is written to the index, blobs or sinks. Remote sinks send metadata only unless you opt in to full content.

What the hook writes to the spool

The hook first writes the agent's payload unredacted to ~/.coding.sh/spool (mode 600). Ingest redacts the payloads kept in spool/<agent>/raw/ in place. A payload written inside its spool line stays unredacted until coding.sh gc deletes the file after 7 days. Details in docs/guides/privacy.md.

Status

coding.sh is pre-1.0 (latest: 0.1.6), so expect rough edges. What isn't done yet:

  • Windows. CI runs on Linux, macOS and Windows. On Windows, CI runs the test suite, the installer (from Git Bash) and real OpenCode, Pi, Goose, Factory Droid, GitHub Copilot CLI and Aider sessions against a stub model. Capture from Claude Code, Codex CLI, Antigravity CLI, Cursor CLI, Amp and T3 Code has not been checked on Windows yet.
  • Not run live. Two adapters, Cursor CLI and Amp, have not been run on a live install yet. Both need an account login (see Supported agents).

Changes are listed in CHANGELOG.md and on the releases page.

Getting help

ForGo to
Questions and ideasDiscussions
BugsOpen an issue. The bug form asks for the output of coding.sh doctor, which answers most first questions.
A new agent, or a report on one not run live yetThe agent support form
Security issuesReport them privately through SECURITY.md, not as an issue

Docs

  • coding.sh/docs: quickstart, agents, CLI, MCP, evals, OpenTelemetry, storage and the trace format.
  • docs/: the same guides in Markdown, plus production, design specs and research.
  • docs/troubleshooting.md: what to check when sessions don't show up.

Repository layout

PathWhat it holds
packages/coding.shThe CLI, hook, adapters, OTLP receiver, viewer, MCP server and evals
packages/sdk, packages/sdk-pythonThe TypeScript and Python SDKs for API apps
packages/store-postgres, packages/store-clickhouse, packages/blob-s3Store plugins
apps/siteThe coding.sh website (Astro, static)
docsGuides, production, reference and design docs
examplesAn eve agent, production apps, a plugin A/B and a skill eval
deployDeploy recipes for Compose, Fly, Railway and ClickHouse (Render uses render.yaml)
evalscoding.sh's own CLI eval suite
scripts/dogfoodThe scorecard coding.sh runs on itself

Contributing

See CONTRIBUTING.md for setup and the codebase map. Issues labeled good first issue are small and self-contained. Release steps are in docs/RELEASE.md.

⭐️ Star the Repository

A drawing of the jackulau/coding.sh repository row. A pointer moves to the Star button and presses it. The star turns yellow and the button reads Starred.

If coding.sh is useful to you, star the repository. A star helps other people find the project.

Star history for jackulau/coding.sh

License

MIT

// faq

What is coding.sh?

Open-source, local-first observability and evals for AI agents: see what Claude Code, Codex, Cursor and others did and what it cost.. It is open-source on GitHub.

Is coding.sh free to use?

coding.sh is open-source under the MIT license, so it is free to use.

What category does coding.sh belong to?

coding.sh is listed under mcp-servers in the Claudeers registry of Claude-compatible tools.

2 views
★ 18 stars
unclaimed
updated 5 days ago

// embed badge

coding.sh on Claudeers
[![Claudeers](https://claudeers.com/api/badge/codingsh.svg)](https://claudeers.com/codingsh)

// retro hit counter

coding.sh hit counter
[![Hits](https://claudeers.com/api/counter/codingsh.svg)](https://claudeers.com/codingsh)

// reviews

// guestbook

0/500

// related in MCP Servers

🔓

f.k.a. Awesome ChatGPT Prompts. Share, discover, and collect prompts from the community. Free and open source — self-host for your organization with complete…

// mcp-serversf/⟨HTML⟩★ 177,810◷ NOASSERTION[ claude ]
🔓

A cross-platform desktop All-in-One assistant for Claude Code, Codex, OpenCode, OpenClaw, Gemini CLI & Hermes Agent. Only official website: ccswitch.io

// mcp-serversfarion1231/⟨Rust⟩★ 155,814◷ MIT[ claude ]
🔓

🪨 why use many token when few token do trick — Claude Code skill that cuts 65% of tokens by talking like caveman

// mcp-serversJuliusBrussee/⟨JavaScript⟩★ 117,780◷ MIT[ claude ]
🔓

An open-source AI agent that brings the power of Gemini directly into your terminal.

// mcp-serversgoogle-gemini/⟨TypeScript⟩★ 108,744◷ Apache-2.0[ claude ]
→ see how coding.sh connects across the ecosystem