Automation

Ships a commit skill and holds every Bash git commit to it: a commit the skill did not open, a blanket git add, a signature trailer, a secret, an ignored path, or a push or branch change nobody asked for is stopped before git runs.

After an edit that calls translation keys, names the keys the JSON, Laravel PHP, YAML or gettext locale files lack, and the languages that lack them.

After each commit the model makes, names the manifests whose dependencies it changed without their lockfile (npm, Composer, Cargo, Go, Python, Bundler, Dart, Mix).

Tells you when an MCP server failed to connect or dropped, with the engine's reason and a reconnect button, and when it is back; claude.ai connectors are left out.

Names each installed plugin, of every marketplace, whose local clone already offers a newer version, with the claude plugin update command that closes the gap.

Lists the servers a Bash call of the model started in this repository and left listening on a port, with their age and session, and stops one with SIGTERM, then SIGKILL.

Measures how much context each Bash command spent and tells the model, past a size limit, which narrower command would have answered the same question.

Keeps the notes you pin with /pin-note for the whole session and sends them to the model again, word for word, after each compaction and /clear. Off until /pin-note on.

For mod development: runs a live check of plugins in a fresh Claude Code session in tmux, types the steps one by one, collects the pane and the transcript, then deletes the temp directory, its transcripts and the inline store files.

Writes a prompt longer than the limit (2000 characters by default) to a temp file and sends the model its first 200 characters with the path, so one paste does not fill the context.

For mod development: lets the model run a slash command in its own session, such as /reload-plugins after a plugin update, and read the output as the next prompt.

Runs slash commands joined with && one after another, waits for each command's turn, pane or dialog to end, and stops the chain at a command that failed or a turn that ended without an answer.

After an edit that builds SQL by joining or interpolating strings, names each line, so the model passes the values as query parameters.

After an edit that keeps browser data in localStorage or sessionStorage, names each line, so the model stores the data in a cookie instead.

Submits a continue prompt while the task list has unfinished tasks, at most 99 times in a row by default, and turns the task tools on for every model unless CLAUDE_CODE_ENABLE_TODO_TOOLS is set.

Stops the model from repeating a tool call that just failed with the same input, until a file or command has changed something.

After a curl or wget an automated-client filter answered 403 or 429, gives the model the browser User-Agent to retry with, and the two cases where it must not.

AGENTS.md as project instructions, by one option. Under claude-md-or-agents-md a project with no instruction files of its own gets its AGENTS.md files instead: a prompt.context hook asks $.fs.ancestors for every AGENTS.md and .claude/AGENTS.md from the filesystem root down to the working directory and hands them to the engine as project instruction files, which it renders, announces and...

Paints Markdown onto Claude Code's prompt box as you type: fenced code becomes a syntax-highlighted card (even before you close the fence), and inline code, bold, italic, lists, task lists, quotes, tables, links and headings are styled. Paint only, your text is never changed.

Client-side secret/PII redaction hooks: keep API keys and sensitive identifiers out of Claude Code's model context by redacting Read/Bash/Grep tool output, blocking prompts that carry secrets, and denying reads of known credential files. Session-transcript scrubbing verified on Claude Code 2.1.263 (updatedToolOutput is what gets persisted); re-verify on other versions.

Compacts the main conversation once it has sat idle for idleMinutes, while the prompt cache is still warm, and hands every compaction of the main conversation (manual, auto, plugin, precompute) the Compact Instructions section of the user's AGENTS.md. Skips while a turn runs, below minMessageTokens of conversation, and past cutoffMinutes (a wake from sleep, when the cache has likely expired)...

Route work through the smallest sufficient loop: handle bounded tasks directly, fan out clear branches, or coordinate durable app-rooted Claude and Codex workrooms with provider-neutral status and git lifecycle tracking.

Reports the install-doctor verdict at session start and refuses tool calls while the install is provably broken

Reports at session start when plugins declared in settings are not actually installed or are disabled for this project