Skip to content

CowboyLogic/ai-dev / skills/client-config-copilotcli

client-config-copilotcli

Unverified

Manage GitHub Copilot CLI configuration files including settings.json, config.json, permissions-config.json, mcp-config.json, lsp-config.json, hooks, skills, custom agents, plugins, and custom instructions. Use this skill whenever the user wants to view, edit, add, or understand any Copilot CLI setting: trusted folders, user or repository settings, tool permissions, sandboxing, MCP or LSP servers, hooks, skills, custom agents, plugins and marketplaces, BYOK models, authentication, session management, or custom instructions. Trigger on: "add trusted folder", "allow tool", "configure MCP", "add hook", "create skill", "create agent", "install plugin", "set model", "use my own API key", "add custom instructions", "name session", "resume session", or "show my copilot config".

FileKindSize
CHANGELOG.mdmarkdown19.4 KB
README.mdmarkdown2.3 KB
SKILL.mdskill md7.6 KB
references/agents-plugins.mdmarkdown11.5 KB
references/cli-commands.mdmarkdown15.2 KB
references/config-schema.mdmarkdown37.2 KB
references/hooks.mdmarkdown15.6 KB
references/instructions.mdmarkdown4.1 KB
references/mcp.mdmarkdown12.1 KB
references/permissions.mdmarkdown22.8 KB
references/skills.mdmarkdown5.8 KB
scripts/show-config.pyscript14.5 KB
scripts/update-references.pyscript5.6 KB
sources.jsonconfig12.2 KB

references/cli-commands.md

View on GitHub
# CLI Commands, Options, and Slash Commands

Run `copilot help [TOPIC]` for built-in help (topics: `billing`, `config`, `commands`,
`environment`, `logging`, `monitoring`, `permissions`, `providers`, `sandbox`). Run `/help` in a
session for the full slash-command list.

## Terminal commands

| Command | Purpose |
|---------|---------|
| `copilot` | Interactive UI |
| `copilot login [--host H] [--web-flow\|--device-code\|--with-token]` | Authenticate |
| `copilot init` | Generate or improve `.github/copilot-instructions.md` |
| `copilot config [KEY] [VALUE]` | Read or change settings outside a session (see below) |
| `copilot sandbox ca status\|create\|trust [CA.PEM]\|rotate\|remove` | Manage the sandbox proxy certificate authority without a session (see `permissions.md`) |
| `copilot workflow run WORKFLOW-NAME [--args JSON\|@PATH] [--result-file PATH] [-s] [--output-format text\|json]` | Run a dynamic workflow registered by an extension, without a chat prompt; can't be combined with `-p`, `-i`, `--agent`, `--fleet`, `--autopilot`, `--resume`, or `--continue` |
| `copilot mcp list\|get\|add\|enable\|disable\|remove` | Manage MCP servers (see `mcp.md`) |
| `copilot skill list\|add\|remove\|enable\|disable` | Manage skills (see `skills.md`) |
| `copilot plugin ...` | Manage plugins and marketplaces (`copilot plugins` is a legacy alias; see `agents-plugins.md`) |
| `copilot instruction list [--json]` | List discovered instruction sources |
| `copilot lsp list [--json]` | List configured language servers |
| `copilot completion bash\|zsh\|fish` | Shell completion script |
| `copilot help [TOPIC]` | Help |
| `copilot update` / `copilot version` | Update / version |
| `copilot app` | Open the GitHub Copilot app in the current directory |

Custom agents and session-scoped hooks have no terminal subcommand — they need a live session.

### `copilot config`

Non-interactive counterpart to `/settings`. Keys use dot notation.

```bash
copilot config theme                       # print a setting
copilot config footer.showQuota off        # set a nested value
copilot config allowedUrls github.com      # append to a list (only if absent)
copilot config --rm allowedUrls github.com # remove one list item (--rm KEY removes the key)
copilot config --repo --list --json        # list repo-overridable settings as JSON
```

| Option | Purpose |
|--------|---------|
| `--list` | List `key=value` lines (conflicts with `KEY`/`VALUE`/`--rm`) |
| `--json` | JSON output for `--list` |
| `--rm` | Remove the key or one list item; requires `KEY` |
| `--global` (default) / `--repo` / `--local` | User `settings.json`, `.github/copilot/settings.json`, or `.github/copilot/settings.local.json`; `--repo` and `--local` accept only repo-overridable settings |

Booleans take `on`/`off` or `true`/`false`. An unset key exits `1`; secret values are redacted. Put
`--` before a value that starts with a dash. Run `copilot help config` for the available settings.

---

## Supported models

`--model=MODEL` or `COPILOT_MODEL` selects the model; `auto` lets Copilot choose. The command
reference lists these IDs:

| Model | Best for |
|-------|----------|
| `claude-sonnet-5.5` | General-purpose coding (default) |
| `gpt-5.4` | Complex reasoning tasks |
| `gpt-6.1-sol`, `gpt-6-astra`, `gpt-6-luna` | Recommended, opt-in (not the automatic default) |
| `gpt-5.6-luna` | Default model of the built-in `task` agent; listed on the supported-models page |
| `claude-opus-5.5` | Recommended, high-capability complex tasks |
| `claude-haiku-5.5`, `claude-haiku-4.5` | Fast, lightweight operations |
| `gpt-5.3-codex` | Code-focused tasks |
| `gemini-3.7-flash`, `gemini-3.8-flash` | Fast Google Gemini responses |

The roster and retirement dates change often; the
[supported models](https://docs.github.com/en/copilot/reference/ai-models/supported-models) page is
the authority.

---

## Command-line options

### Session and mode

| Option | Purpose |
|--------|---------|
| `-p PROMPT`, `--prompt=PROMPT` | Run programmatically and exit |
| `-i PROMPT`, `--interactive=PROMPT` | Start interactive and run this prompt |
| `-n NAME`, `--name=NAME` | Name the new session |
| `-r`, `--resume[=VALUE]` | Resume by ID, ID prefix, or name (exact, case-insensitive); bare opens picker (needs TTY) |
| `--continue` | Resume most recent session in cwd (else globally) |
| `--session-id ID` | Exact session/task ID; creates a new session only for a valid UUID |
| `--connect[=SESSION-ID]` | Connect to a remote session |
| `--sync-session[=on\|off]` | Turn read-only session-history syncing on or off for this run. Without a value it turns syncing on. It does not enable remote control; use `/remote on` for that |
| `-C DIRECTORY` | Change directory first |
| `-w`, `--worktree[=NAME]` | Start in a Git worktree under `<repo>.worktrees/` |
| `--mode=interactive\|plan\|autopilot` | Initial mode; `--plan --mode autopilot` = plan-then-autopilot |
| `--plan` | Start in plan mode |
| `--autopilot` | Keep working until `task_complete` |
| `--max-autopilot-continues=COUNT` | Cap autopilot continuations |
| `--fleet` | Run the prompt with parallel subagents |
| `--agent=AGENT` | Use a custom agent |
| `--model=MODEL` | Model (`auto` allowed) |
| `--effort=LEVEL`, `--reasoning-effort=LEVEL` | `low`, `medium`, `high`, `xhigh`, `max` |
| `--context default\|long_context` | Context tier |
| `--max-ai-credits=CREDITS` | Soft per-response AI credit cap |
| `--acp` | Run as Agent Client Protocol server. Sessions honor saved sandbox settings and `--sandbox`/`--no-sandbox`, including on resume; a managed policy that enforces sandboxing overrides `--no-sandbox` |

### Permissions and tools

`--allow-all`, `--yolo`, `--allow-all-tools`, `--allow-all-paths`, `--allow-all-urls`,
`--allow-tool`, `--deny-tool`, `--allow-url`, `--deny-url`, `--available-tools`,
`--excluded-tools`, `--disallow-temp-dir`, `--add-dir`, `--no-ask-user`, `--sandbox`,
`--no-sandbox`, `--secret-env-vars=VAR`, `--trust-folder` — see `permissions.md`.

### MCP, plugins, instructions

| Option | Purpose |
|--------|---------|
| `--additional-mcp-config=JSON\|@FILE` | Add MCP servers for this session (highest priority) |
| `--disable-builtin-mcps` | Disable built-in MCP servers |
| `--disable-mcp-server=NAME` / `--enable-mcp-server=NAME` | Per-session toggle |
| `--add-github-mcp-tool=TOOL` / `--add-github-mcp-toolset=TOOLSET` / `--enable-all-github-mcp-tools` | Widen GitHub MCP tool set |
| `--allow-all-mcp-server-instructions` | Put all MCP server instructions in the system prompt |
| `--mcp-github-auth` | Scope GitHub account auth to approved MCP server origins (v1.0.90 release notes; not in the docs reference — run `copilot --help` for exact semantics) |
| `--plugin-dir=DIRECTORY` | Load a plugin from a local directory |
| `--extension-sdk-path=DIRECTORY` | Override the bundled `@github/copilot-sdk` injected into extension subprocesses (invalid paths fall back to the bundled SDK) |
| `--no-custom-instructions` | Skip `AGENTS.md` and related instruction files |
| `--enable-memory` | Enable memory in prompt mode |
| `--disable-client-telemetry` | Turn off first-party client-side telemetry, unrestricted and restricted, for this invocation only; not saved. It does not affect server-side collection, OpenTelemetry exports you or your enterprise configure, or telemetry an SDK application explicitly forwards from the CLI |

### Output and UI

`--output-format=text|json` (JSONL; v1.0.97 release notes: secrets are redacted in JSON output
while valid variable references are preserved), `-s`/`--silent`, `--stream=on|off`, `--share=PATH`,
`--share-gist`, `--attachment PATH`, `--log-dir`, `--log-level`, `--banner`/`--no-banner`,
`--no-color`, `--plain-diff`, `--screen-reader`, `--mouse[=on|off]`/`--no-mouse`,
`--experimental`/`--no-experimental`, `--no-auto-update`, `--bash-env`/`--no-bash-env`.

### Removed and renamed remote options

Per the command reference, `--no-remote` is removed, `--remote-export` is renamed
`--sync-session=on`, and `--no-remote-export` is renamed `--sync-session=off`. `--remote` is not
supported for interactive startup or with `-p`; use `/remote on` in a session (`/remote off` ends
the connection). It remains supported only for legacy SDK server startup with `--server` or
`--headless`. The release notes for v1.0.103 say the legacy `--remote`/`--no-remote` startup flags
were retired in favor of clearer hosting and connection flags; they don't name the new flags, so
run `copilot --help` for them.

### Deprecated options

| Option | Status |
|--------|--------|
| `--config-dir=DIRECTORY` | Deprecated — use `COPILOT_HOME` |
| `--enable-reasoning-summaries` | Accepted but ignored |

---

## Sessions

```bash
copilot --name my-feature-work      # name a new session
copilot --resume=my-feature-work    # resume by name, ID, or ID prefix
copilot --continue                  # most recent session in this directory
```

| Slash command | Purpose |
|---------------|---------|
| `/session [info\|checkpoints [n]\|files\|plan\|rename [NAME]\|cleanup\|prune\|delete [ID]\|delete-all]` | Session info and management (`/sessions` alias) |
| `/rename [NAME]` | Rename current session |
| `/resume [ID]`, `/continue [ID]` | Switch session |
| `/fork [NAME]`, `/branch [NAME]` | Fork current session |
| `/clear`, `/new`, `/reset` | Start a new conversation |
| `/undo`, `/rewind` | Roll back to an earlier turn (optionally restore files) |
| `/restart` | Restart the CLI, restoring live sessions |
| `/share [link\|off\|file\|html\|gist\|research]`, `/export` | Share/export the session. With no subcommand it opens a sharing menu and shares nothing; `link` shares and prints a GitHub link, `off` (or `link off`) stops link sharing |

`/session cleanup` shows a storage summary of your other local sessions grouped by age and suggests
deletion commands. It deletes nothing; preview with `/session prune --older-than DAYS --dry-run`.

Session history lives in `~/.copilot/session-state/`.

---

## Slash commands by area

### Configuration

| Command | Purpose |
|---------|---------|
| `/settings [--repo\|--local] [show KEY\|KEY\|KEY VALUE]`, `/config` | Settings editor or inline set |
| `/model [--session\|--global\|--repo\|--local] [MODEL\|auto TIER]`, `/models` | Choose model, effort, context, or an Auto tier. Model IDs are case-insensitive; the canonical ID is saved |
| `/theme [default\|github\|dim\|high-contrast\|colorblind]` | Color theme |
| `/statusline`, `/footer` | Status-line items |
| `/experimental [on\|off\|show]` | Experimental features |
| `/terminal-setup` | Multiline input keys |
| `/vim` | Toggle Vim mode |
| `/keep-alive [on\|off\|busy\|DURATION]`, `/caffeinate` | Prevent sleep |
| `/limits`, `/limits set max-ai-credits VALUE`, `/limits unset` | Per-response credit limits |

### Customization

| Command | Purpose |
|---------|---------|
| `/mcp [config\|list\|show\|add\|edit\|delete\|disable\|enable\|auth\|reload\|search] [NAME]` | MCP servers |
| `/skills [list\|info\|add\|remove\|reload]` | Skills |
| `/plugin [install\|update\|uninstall\|list\|marketplace ...]` | Plugins. `list` shows team-assigned plugins that are default-enabled as enabled, unless you disable them or managed settings override the default; plugins enforced through managed `enabledPlugins` can't be disabled locally |
| `/agent` | Pick a custom agent; press `N` in the list to create one |
| `/subagents`, `/agents` | Per-agent subagent models |
| `/instructions` | View and toggle instruction files |
| `/lsp [show\|test\|reload\|logs\|help] [NAME]` | Language servers |
| `/extensions`, `/extension` | Extensions (experimental). `/extensions mode` sets **Load & Augment** (default; extensions run and the agent can manage them), **Load Only** (run, no agent management tools), or **Disabled** (none run; persists for future sessions). `/extensions manage` inspects them |
| `/env` | Show loaded instructions, MCP, skills, agents, hooks, plugins, LSPs, extensions |
| `/init` | Create/update `copilot-instructions.md` (`/init suppress` hides the startup hint) |

The experimental `/plugins` command was removed — use `/plugin`, `/mcp`, `/skills`, `/subagents`,
and `/instructions`.

### Permissions and directories

`/permissions [default|assisted|allow-all|show|reset]`, `/allow-all`, `/yolo`,
`/reset-allowed-tools`, `/add-dir PATH`, `/list-dirs`, `/cwd`, `/cd [PATH]`,
`/sandbox [config|status|policy [COMMAND]|enable|disable]`, `/sandbox enable session`,
`/sandbox disable session`, `/sandbox ca [create|trust|rotate|remove]` (see `permissions.md`).

### Work

| Command | Purpose |
|---------|---------|
| `/plan [PROMPT]` | Plan before coding (project files write-protected) |
| `/autopilot [OBJECTIVE]`, `/goal` | Autopilot, optional `--max-ai-credits N` |
| `/fleet [PROMPT]` | Parallel subagents |
| `/delegate [PROMPT]` | Hand off to Copilot cloud agent (opens a PR) |
| `/review [PROMPT]` / `/security-review [PROMPT]` | Code / security review agents |
| `/rubber-duck [PROMPT]` | Second-opinion agent |
| `/research TOPIC` | Deep research |
| `/pr [view\|create\|fix\|auto\|automerge]` | Pull requests for the current branch |
| `/diff` | Review changes |
| `/worktree [branch\|task]`, `/worktree new`, `/move` | Git worktrees |
| `/every`, `/after` | Scheduled prompts (experimental) |
| `/tasks` | Background subagents and shells |
| `/ask [QUESTION]`, `/btw [QUESTION]` | Side question outside history; the answer renders as Markdown. In the answer dialog, `Ctrl+E` expands it and `Ctrl+Y` copies the full answer. With no question, reopens the last side-question answer in the session |
| `/compact [FOCUS]` / `/context` | Compress / inspect context |
| `/chronicle <standup\|tips\|improve\|reindex\|skills ...>` | Session-history insights |
| `/usage` | Usage and AI credits (v1.0.97 release notes: shows billed AI Credits for token-billed sessions, including after resume) |
| `/diagnose [PROMPT]` | Analyze the session log |
| `/collect-debug-logs [file\|gist\|share] [PATH]` | Collect debug logs into a local archive (`file`, at `PATH` when given), a secret GitHub gist (`gist`), or a local archive that is then uploaded to GitHub (`share`) |
| `/refine TEXT` | Rewrite a rough prompt into a clear one for review (no argument: clean up the input box) |
| `/workflows` | Observe workflow runs, phases, agents, and progress (needs a token-based-billing plan) |

### Account and app

`/login`, `/logout`, `/user [show|list|switch]` (`/user list` shows the authentication source of
each signed-in account, per the v1.0.96 release notes), `/remote [on|off]`, `/ide`, `/app`,
`/feedback` (`/bug`), `/changelog [summarize] [VERSION|last N|since VERSION]` (`/release-notes`),
`/update` (`/upgrade`), `/version`, `/voice`, `/copy`, `/search` (`/find`),
`/downgrade VERSION` (team accounts), `/collect-debug-logs` (see Debug logs below),
`/clikit [COMPONENT]` (preview CLI business components), `/tuikit [colors|icons|select|tabbar]` (preview TUIkit design-system components),
`/exit` (`/quit`).

### Debug logs

`/collect-debug-logs` and the feedback-with-logs route offered from `/feedback` first save a local
archive. `share` and the feedback route then ask for fresh consent, defaulting to **No**, before
uploading to GitHub. Per the docs, uploading is currently available only to GitHub staff and
internal users; everyone else uses the `file` and `gist` routes and the existing public feedback
route.

OpenTelemetry export is documented in `config-schema.md` §OpenTelemetry monitoring.
reverse-engineer-anythingmorluto/reaReverse engineer native, managed, Electron/JavaScript, packaged, firmware, and browser targets with REA, and build agent environments (AgentEnv bundles) of web applications from a live instance or a capture, and of Linux desktop apps with a task from a request. Use shipped-artifact or requested runtime evidence to explain features, compare versions, decompile code, or guide a reconstruction. Skip REA for ordinary source-repository architecture analysis.Stars: 468kUpdated Claude CodeCodexVerifiedInstall
diagnosing-superpowersobra/superpowersUse when a superpowers session went wrong and your human partner wants to know why — repeated work, ignored plans, stumbles, poor results, a skill that didn't fire, "it took too long", "why is it so expensive", "what is it doing" — or wants to build a bug report for the superpowers maintainers, for the current session or a past one identified by id or path, on any harness.Stars: 321kUpdated Claude CodeCodexVerifiedInstall
dispatching-parallel-agentsobra/superpowersUse when facing 2+ independent tasks that can be worked on without shared state or sequential dependenciesStars: 321kUpdated Claude CodeCodexVerifiedInstall
executing-plansobra/superpowersUse when executing an implementation plan in the current session as the implementer yourself — your human partner chose inline execution, or no subagent tool is availableStars: 321kUpdated Claude CodeCodexVerifiedInstall