INTERNAL reference, NOT shipped with the installer. The in-repo expert on every coding-agent CLI CC Director can run (Claude Code, Codex, Gemini, OpenCode, Grok, Copilot, Cursor, Pi). Knows each agent's command line, config files, context-injection points, lifecycle hooks and events, SDK/API/server mode, transcript format, session semantics, and how CC Director drives it. Triggers on "agent-expert", "agent expert", "how does <agent> inject context", "<agent> hooks", "<agent> events", "<agent> CLI reference", "which agents support X", "agent capability matrix".

RETIRED - the Director's Control API no longer exists (the remove-the-network-port mission deleted the listener), so settings can no longer be driven over local HTTP. Kept only to say where settings editing lives now.

A check whose pass condition is an ABSENCE certifies a run that never happened. Restate it as a specific PRESENCE - and an empty result is a broken instrument, never a clean run. Use when the task involves: verify, prove it, sweep, grep returned nothing, zero hits, clean run, it passed, no defects, reviewed, delivered, double check, how do we know.

Audit the cc-director repo for hygiene problems (junk files, root clutter, stale docs, dead/unreferenced code, archived dirs) and produce an incremental cleanup plan, optionally filing one-at-a-time GitHub cleanup issues. Triggers on "/code-maintenance", "clean up the repo", "repo hygiene", "code maintenance", "tidy the codebase".

A destructive operation acts only on what it can positively prove is disposable. Enumerate what to DELETE, never what to skip - and refresh the safety signal on every real use. Use when the task involves: sweep, purge, cleanup, delete old, prune, reclaim space, stale files, retention, aged out, rm -rf, git clean, drop the table.

**When the owner asks you for a report, a write-up, a mission document, a design, findings, or "tell me what you found" - you publish a DEV REPORT. One HTML file, published with `cc-dev-reports`. That is the default and it needs no permission.**

DevThrottle is a desktop application positioned as "Mission Control for Claude Code" - one place to run, observe, and orchestrate multiple Claude Code sessions side by side. It also installs a suite of `cc-*` command-line tools onto your PATH.

One word per idea, one idea per word. These are the definitions everything else in DevThrottle uses - the product, the documentation, the workflow conduct, and every agent working in this fleet. If you are writing an issue, a brief, a commit message, a document, a code comment, or a message to anoth

Relocate work from one Director to another - a newer build, a different slot, another machine.

Three rules. They take ten seconds each at spawn time and cannot be recovered cheaply afterwards.

A proof can be valid and still say nothing about what you changed - wrong surface, wrong reader, wrong caller, or an untested premise. Name what your evidence does NOT cover. Use when the task involves: prove it, by construction, root cause, reproduced it, repro, guard-local, skipped test, suite is green, premise, assumption, self review, it is settled.

The Support Agent in the CenCon Development Method for cc-director. The idle question-answering seat - answers questions about the cc-director codebase and about what the Product, Developer, and QA agents are doing, and owns/maintains the CenCon documents. Strictly read-only with respect to product code and issues - it never implements, never creates or moves issues. Triggers on "/support-agent", "support agent", "ask about the codebase", "what are the agents doing", "explain how X works".

How we build software here. Version 1, 18 September 2026. Read your own seat's section - it answers everything you need, including about the seats around you.

Interact with and test the CC Director WPF app using the cc-click CLI tool. Use this to click buttons, read text, take screenshots, list elements, right-click session tabs, and navigate the UI.

Review the voice-turn comparison log for fidelity drift (what the user said vs what was cleaned, and what the agent said vs what the wingman spoke), archive flagged turns locally, and file sanitized GitHub issues. Triggers on "/voice-review", "review voice log", "review wingman fidelity", "review voice turns".