Claude Code Mods Overview Create Claude Code Mods

• agent-profiler: shows where a session's time goes (model wait, tools, subagents, re-reads) and exports OTLP traces for Jaeger/Tempo.

• context-inspector (한국어): analyzes context waste in Claude sessions, including what fills the window, re-reads, near-duplicate outputs, and what compaction kept, providing advice.

• time-machine (한국어): records every tool call on a branching timeline with shadow-git snapshots, so you can see the files as they were at any step, explain which edit broke a test, restore, branch, compare runs and re-run steps.

• loop-detective (한국어): detects repeated Bash failures and edit/retry cycles, with evidence, investigation prompts, a live case pane and bounded history.

• pin-memo (한국어): pins memos and clickable todos above the input, restoring them when the same folder/session is resumed.

• model-effort-band (한국어): shows the selected model and observed main-thread effort in a colored band above the prompt.

• token-run (한국어): a fully local ASCII runner for request failures and short breaks, with jump controls and a saved best score.

• verify-board (한국어): Revision-aware verification results: PASS, FAIL, STALE and UNKNOWN.

• scope-guard (한국어): Allowed-path warnings with rename and pre-existing-change awareness.

• blast-radius (한국어): Local reverse-import impact and related test candidates.

• decision-ledger (한국어): Explicit decisions with rationale, rejected alternatives and file references.

• handoff-pack (한국어): Resumable Markdown previews of goals, tasks, changes and evidence.

• command-shelf (한국어): Successful-command history with explicit permission-aware replay.

• env-doctor (한국어): Project runtime requirements and environment-variable presence checks.

• dependency-lens (한국어): Direct and transitive npm dependency changes and install-script markers.

• assumption-tracker (한국어): Evidence-backed assumption states with retained transition history.

• review-queue (한국어): Purpose-grouped file review with staged/unstaged freshness checks.

The ten new workflow mods can be loaded from this checkout before the branch is published to the marketplace. Choose a mod directory, for example:

claude --plugin-dir ./verify-board

Commands: /verify, /scope, /blast, /decision, /handoff, /shelf, /env-doctor, /deps, /assume, /review-queue. Each accepts help. These first releases return command text; they do not add live panels.

Use a current Claude Code CLI with Mods support (mods are enabled by default from 2.1.287). Token Run has been verified in real terminal sessions with 2.1.296 and 2.1.301; see each mod's README for its tested versions. Time Machine also requires git on your PATH.

Run this entire block in your terminal (macOS/Linux, Bash or Zsh). It adds the marketplace and installs all seventeen mods for your user account, so they are available across projects. No repository clone or npm install is needed.

(
  set -eu
  claude plugin marketplace add YeonwooSung/my-claude-code-mods

  for mod in agent-profiler context-inspector time-machine loop-detective pin-memo model-effort-band token-run verify-board scope-guard blast-radius decision-ledger handoff-pack command-shelf env-doctor dependency-lens assumption-tracker review-queue; do
    claude plugin install "${mod}@my-claude-code-mods" --scope user
  done
)

If you have already added the marketplace, omit the claude plugin marketplace add line. The block stops on the first failed command; resolve that error before installing the remaining mods.

Add the marketplace once in your terminal:

claude plugin marketplace add YeonwooSung/my-claude-code-mods

Then run only the install commands for the mods you want:

claude plugin install agent-profiler@my-claude-code-mods --scope user
claude plugin install context-inspector@my-claude-code-mods --scope user
claude plugin install time-machine@my-claude-code-mods --scope user
claude plugin install loop-detective@my-claude-code-mods --scope user
claude plugin install pin-memo@my-claude-code-mods --scope user
claude plugin install model-effort-band@my-claude-code-mods --scope user
claude plugin install token-run@my-claude-code-mods --scope user
claude plugin install verify-board@my-claude-code-mods --scope user
claude plugin install scope-guard@my-claude-code-mods --scope user
claude plugin install blast-radius@my-claude-code-mods --scope user
claude plugin install decision-ledger@my-claude-code-mods --scope user
claude plugin install handoff-pack@my-claude-code-mods --scope user
claude plugin install command-shelf@my-claude-code-mods --scope user
claude plugin install env-doctor@my-claude-code-mods --scope user
claude plugin install dependency-lens@my-claude-code-mods --scope user
claude plugin install assumption-tracker@my-claude-code-mods --scope user
claude plugin install review-queue@my-claude-code-mods --scope user

Alternatively, run these slash commands inside a Claude Code session, replacing pin-memo with any mod name listed above:

/plugin marketplace add YeonwooSung/my-claude-code-mods
/plugin install pin-memo@my-claude-code-mods

The shell commands above explicitly use --scope user. For a project-specific installation, use --scope project (shared through the project's settings) or --scope local (only you in that project).

Start a new Claude Code session after installation, or run /reload-plugins in an already open session. Run /plugin to check the loaded mods and manage or disable individual ones. See the official Mods installation documentation and each mod's README under Mods for commands and configuration.

Expanded workflow validation (2026-10-18) documents 55 added tests, real-Git regressions and ten-adapter controlled-host composition. Individual engine suites are verified; live UI and actual combined engine loading require separate integration checks.

Use Node 22.19.0 or later. Dependencies are pinned in package-lock.json.

npm ci --ignore-scripts
npm test                  # core logic, no Claude Code installation/account required
npm run test:privacy      # shared masking corpus and real-Git privacy regressions
CLAUDE_BIN=/path/to/claude npm run test:adapters  # full engine suites, including adapters
npm run check:workflow    # verify standalone runtime copies match canonical sources
npm run test:workflow     # real-Git and queued-session regression tests
npm run typecheck         # requires engine-generated declarations in each .claude-plugin/types

npm test maps the core tests' claude-code/testing import to Node's test runner and Jest's expect; it does not emulate engine events. Files run sequentially so the existing performance assertions do not compete with other test processes. CI runs this independent core suite. Engine type declarations are generated locally and remain untracked; use the declarations belonging to the engine under test. See each mod's development instructions for generation/setup.

For a controlled overhead comparison, use a disposable representative large repository at one fixed commit and the same Claude engine/model/permission settings. Run the same workload under four configurations: no mods, profiler only, Time Machine only, and all three. Load mods explicitly with --plugin-dir while other installed plugins are disabled; reset files and recorder stores outside the timed interval. Use a warm-up and at least five measured repetitions per configuration, alternate run order, and record wall time, tool-call count, model tokens and snapshot average. Include no-change mutating calls and actual edits. Report median/p95 and separate model latency from snapshot/hook time. The existing 20k-span/item tests bound analysis cost but do not measure combined engine overhead; no four-mode performance result is claimed here.