feat(copilot): add context and reasoning_effort config fields

CHANGELOG.md

@@ -10,7 +10,7 @@ Adds HTTP request-body inspection to Keep policies. File- and pack-based `networ
 
 ### Added
 
-- **GitHub Copilot CLI agent** — run GitHub Copilot CLI with `moat copilot`. Copilot uses the existing `github` grant: Moat injects that GitHub token for GitHub/Copilot API hosts plus HTTPS git, while the container receives only placeholders. `moat copilot` installs `@github/copilot`, stages Copilot config/context, passes `--allow-all` by default, and supports `copilot.model`, `copilot.experimental`, and `copilot.autopilot` in `moat.yaml`. See [Running GitHub Copilot CLI](https://majorcontext.com/moat/guides/copilot). ([#436](https://github.com/majorcontext/moat/pull/436))
+- **GitHub Copilot CLI agent** — run GitHub Copilot CLI with `moat copilot`. Copilot uses the existing `github` grant: Moat injects that GitHub token for GitHub/Copilot API hosts plus HTTPS git, while the container receives only placeholders. `moat copilot` installs `@github/copilot`, stages Copilot config/context, passes `--allow-all` by default, and supports `copilot.model`, `copilot.context`, `copilot.reasoning_effort`, `copilot.experimental`, and `copilot.autopilot` in `moat.yaml`. See [Running GitHub Copilot CLI](https://majorcontext.com/moat/guides/copilot). ([#436](https://github.com/majorcontext/moat/pull/436))
 - **Pi packages & safe defaults** — declare Pi extensions/skills/themes in `pi.packages` (remote `npm:`/`git:`/`https:`/`ssh:` sources) and Moat installs them into the image at build time via `pi install`, baked into a reproducible cached layer. Every `moat pi` image also bakes a safe `~/.pi/agent/settings.json` — `defaultProjectTrust: never` (a checked-out repo's own `.pi/` extensions, which are arbitrary code, do not auto-load), telemetry off, quiet startup — that a workspace cannot override. Because Pi config can redirect model traffic to any host, `moat pi` now warns under a permissive network policy (only `network.policy: strict` truly constrains egress). See [Running Pi](https://majorcontext.com/moat/guides/pi). ([#434](https://github.com/majorcontext/moat/pull/434))
 - **Pi coding agent** — run the [Pi coding agent](https://github.com/earendil-works/pi) with `moat pi`. Pi has no credential of its own; it runs against your existing `anthropic` or `openai` grant. When exactly one is configured it is used automatically; when both are, choose one with `--provider` or `pi.provider` in `moat.yaml`. Only the `anthropic` and `openai` backends are supported today — any other backend, or a missing/ambiguous grant, fails before a container is created. Configure with the `pi:` block (`provider`, `model`). See [Running Pi](https://majorcontext.com/moat/guides/pi) and `examples/agent-pi`. ([#433](https://github.com/majorcontext/moat/pull/433))
 - **`opentofu` and `terragrunt` dependencies** — two new managed cloud tools. `opentofu` installs the OpenTofu CLI as the `tofu` command; `terragrunt` installs the Terragrunt orchestration wrapper. Both install as prebuilt release binaries with no image rebuild cost beyond their own layer. Terragrunt delegates to a Terraform or OpenTofu binary on `PATH`, so pair it with an engine — `dependencies: [terraform, terragrunt]`, or `dependencies: [opentofu, terragrunt]` with `env.TERRAGRUNT_TFPATH: tofu`. See [Dependencies](https://majorcontext.com/moat/reference/dependencies). ([#430](https://github.com/majorcontext/moat/pull/430))

docs/content/guides/17-copilot.md

@@ -65,6 +65,8 @@ grants:
 
 copilot:
   model: gpt-5.4
+  context: long_context
+  reasoning_effort: high
   experimental: false
   autopilot: false
 ```

docs/content/reference/01-cli.md

@@ -308,6 +308,8 @@ In addition to the command-specific flags below, `moat copilot` accepts all [com
 | `-p`, `--prompt TEXT` | Run non-interactive with prompt |
 | `--allow-all` | Allow all Copilot tools, paths, and URLs without prompting. Default: `true`. |
 | `--model MODEL` | Select the model to use. Overrides `copilot.model`. |
+| `--context TIER` | Set context window tier (`default` or `long_context`). Overrides `copilot.context`. |
+| `--reasoning-effort LEVEL` | Set reasoning effort level (`none`, `minimal`, `low`, `medium`, `high`, `xhigh`, `max`). Overrides `copilot.reasoning_effort`. |
 | `--experimental` | Enable Copilot CLI experimental features. |
 | `--autopilot` | Start Copilot CLI in autopilot mode. |
 

docs/content/reference/02-moat-yaml.md

@@ -129,6 +129,8 @@ codex:
 # GitHub Copilot CLI
 copilot:
   model: gpt-5.4
+  context: long_context
+  reasoning_effort: high
   experimental: false
   autopilot: false
 
@@ -1595,6 +1597,34 @@ copilot:
 - Default: Copilot CLI default
 - CLI override: `moat copilot --model`
 
+### copilot.context
+
+Context window tier to pass to `copilot --context`.
+
+```yaml
+copilot:
+  context: long_context
+```
+
+- Type: `string`
+- Values: `default`, `long_context`
+- Default: Copilot CLI default
+- CLI override: `moat copilot --context`
+
+### copilot.reasoning_effort
+
+Reasoning effort level to pass to `copilot --reasoning-effort`.
+
+```yaml
+copilot:
+  reasoning_effort: high
+```
+
+- Type: `string`
+- Values: `none`, `minimal`, `low`, `medium`, `high`, `xhigh`, `max`
+- Default: Copilot CLI default
+- CLI override: `moat copilot --reasoning-effort`
+
 ### copilot.experimental
 
 Start Copilot CLI with experimental features enabled.

internal/config/config.go

@@ -339,6 +339,14 @@ type CopilotConfig struct {
 	// Model optionally selects the model passed to `copilot --model`.
 	Model string `yaml:"model,omitempty"`
 
+	// Context sets the context window tier passed to `copilot --context`.
+	// Valid values: "default", "long_context".
+	Context string `yaml:"context,omitempty"`
+
+	// ReasoningEffort sets the reasoning effort level passed to `copilot --reasoning-effort`.
+	// Valid values: "none", "minimal", "low", "medium", "high", "xhigh", "max".
+	ReasoningEffort string `yaml:"reasoning_effort,omitempty"`
+
 	// Experimental starts Copilot CLI with --experimental.
 	Experimental bool `yaml:"experimental,omitempty"`
 

internal/providers/copilot/cli.go

@@ -18,9 +18,13 @@ var (
 	copilotWtFlag               string
 	copilotAllowAll             bool
 	copilotModelFlag            string
+	copilotContextFlag          string
+	copilotReasoningEffortFlag  string
 	copilotExperimental         bool
 	copilotAutopilot            bool
 	copilotResolvedModel        string
+	copilotResolvedContext      string
+	copilotResolvedEffort       string
 	copilotCredentialConfigured = defaultCopilotCredentialConfigured
 )
 
@@ -65,6 +69,8 @@ Examples:
 	copilotCmd.Flags().StringSliceVar(&copilotAllowedHosts, "allow-host", nil, "additional hosts to allow network access to")
 	copilotCmd.Flags().BoolVar(&copilotAllowAll, "allow-all", true, "allow all Copilot tools, paths, and URLs without prompting")
 	copilotCmd.Flags().StringVar(&copilotModelFlag, "model", "", "model to use (overrides copilot.model)")
+	copilotCmd.Flags().StringVar(&copilotContextFlag, "context", "", "context window tier: default or long_context (overrides copilot.context)")
+	copilotCmd.Flags().StringVar(&copilotReasoningEffortFlag, "reasoning-effort", "", "reasoning effort level (overrides copilot.reasoning_effort)")
 	copilotCmd.Flags().BoolVar(&copilotExperimental, "experimental", false, "enable Copilot CLI experimental features")
 	copilotCmd.Flags().BoolVar(&copilotAutopilot, "autopilot", false, "start Copilot CLI in autopilot mode")
 	copilotCmd.Flags().StringVar(&copilotWtFlag, "worktree", "", "run in a git worktree for this branch")
@@ -98,11 +104,19 @@ func runCopilot(cmd *cobra.Command, args []string) error {
 
 func resolveCopilotPreflight(cfg *config.Config) error {
 	copilotResolvedModel = copilotModelFlag
+	copilotResolvedContext = copilotContextFlag
+	copilotResolvedEffort = copilotReasoningEffortFlag
 	if cfg != nil {
 		cfg.Grants = normalizeCopilotGrants(cfg.Grants, false)
 		if copilotResolvedModel == "" {
 			copilotResolvedModel = cfg.Copilot.Model
 		}
+		if copilotResolvedContext == "" {
+			copilotResolvedContext = cfg.Copilot.Context
+		}
+		if copilotResolvedEffort == "" {
+			copilotResolvedEffort = cfg.Copilot.ReasoningEffort
+		}
 		if cfg.Copilot.Experimental {
 			copilotExperimental = true
 		}
@@ -167,6 +181,12 @@ func buildCopilotCommand(promptFlag, initialPrompt string) []string {
 	if copilotResolvedModel != "" {
 		cmd = append(cmd, "--model", copilotResolvedModel)
 	}
+	if copilotResolvedContext != "" {
+		cmd = append(cmd, "--context", copilotResolvedContext)
+	}
+	if copilotResolvedEffort != "" {
+		cmd = append(cmd, "--reasoning-effort", copilotResolvedEffort)
+	}
 	if copilotExperimental {
 		cmd = append(cmd, "--experimental")
 	}

internal/providers/copilot/provider_test.go

@@ -131,7 +131,7 @@ func TestRegisterCLI(t *testing.T) {
 	if cmd == nil || cmd.Use != "copilot [workspace] [flags] [-- initial-prompt]" {
 		t.Fatalf("registered command = %#v", cmd)
 	}
-	for _, flag := range []string{"prompt", "allow-all", "model", "experimental", "autopilot", "worktree"} {
+	for _, flag := range []string{"prompt", "allow-all", "model", "context", "reasoning-effort", "experimental", "autopilot", "worktree"} {
 		if cmd.Flags().Lookup(flag) == nil {
 			t.Fatalf("copilot command missing --%s flag", flag)
 		}
@@ -140,13 +140,19 @@ func TestRegisterCLI(t *testing.T) {
 
 func TestResolveCopilotPreflight(t *testing.T) {
 	origModelFlag, origResolvedModel := copilotModelFlag, copilotResolvedModel
+	origContextFlag, origResolvedContext := copilotContextFlag, copilotResolvedContext
+	origEffortFlag, origResolvedEffort := copilotReasoningEffortFlag, copilotResolvedEffort
 	origExperimental, origAutopilot := copilotExperimental, copilotAutopilot
 	origFlagGrants := copilotFlags.Grants
 	origDryRun := cli.DryRun
 	origConfigured := copilotCredentialConfigured
 	t.Cleanup(func() {
 		copilotModelFlag = origModelFlag
 		copilotResolvedModel = origResolvedModel
+		copilotContextFlag = origContextFlag
+		copilotResolvedContext = origResolvedContext
+		copilotReasoningEffortFlag = origEffortFlag
+		copilotResolvedEffort = origResolvedEffort
 		copilotExperimental = origExperimental
 		copilotAutopilot = origAutopilot
 		copilotFlags.Grants = origFlagGrants
@@ -155,14 +161,22 @@ func TestResolveCopilotPreflight(t *testing.T) {
 	})
 
 	copilotModelFlag = ""
+	copilotContextFlag = ""
+	copilotReasoningEffortFlag = ""
 	copilotExperimental = false
 	copilotAutopilot = false
 	copilotFlags.Grants = []string{"copilot", "ssh:github.com"}
 	cli.DryRun = false
 	copilotCredentialConfigured = func() bool { return false }
 	cfg := &config.Config{
-		Grants:  []string{"github", "copilot"},
-		Copilot: config.CopilotConfig{Model: "gpt-5.4", Experimental: true, Autopilot: true},
+		Grants: []string{"github", "copilot"},
+		Copilot: config.CopilotConfig{
+			Model:           "gpt-5.4",
+			Context:         "long_context",
+			ReasoningEffort: "high",
+			Experimental:    true,
+			Autopilot:       true,
+		},
 	}
 
 	if err := resolveCopilotPreflight(cfg); err != nil {
@@ -171,6 +185,12 @@ func TestResolveCopilotPreflight(t *testing.T) {
 	if copilotResolvedModel != "gpt-5.4" || !copilotExperimental || !copilotAutopilot {
 		t.Fatalf("resolved state = model:%q experimental:%v autopilot:%v", copilotResolvedModel, copilotExperimental, copilotAutopilot)
 	}
+	if copilotResolvedContext != "long_context" {
+		t.Fatalf("resolved context = %q, want long_context", copilotResolvedContext)
+	}
+	if copilotResolvedEffort != "high" {
+		t.Fatalf("resolved effort = %q, want high", copilotResolvedEffort)
+	}
 	if !slices.Equal(cfg.Grants, []string{"github"}) {
 		t.Fatalf("config grants = %v, want [github]", cfg.Grants)
 	}
@@ -179,22 +199,34 @@ func TestResolveCopilotPreflight(t *testing.T) {
 	}
 }
 
-func TestResolveCopilotPreflightModelFlagOverridesConfig(t *testing.T) {
+func TestResolveCopilotPreflightFlagOverridesConfig(t *testing.T) {
 	origModelFlag, origResolvedModel := copilotModelFlag, copilotResolvedModel
+	origContextFlag, origResolvedContext := copilotContextFlag, copilotResolvedContext
+	origEffortFlag, origResolvedEffort := copilotReasoningEffortFlag, copilotResolvedEffort
 	origDryRun := cli.DryRun
 	origConfigured := copilotCredentialConfigured
 	t.Cleanup(func() {
 		copilotModelFlag = origModelFlag
 		copilotResolvedModel = origResolvedModel
+		copilotContextFlag = origContextFlag
+		copilotResolvedContext = origResolvedContext
+		copilotReasoningEffortFlag = origEffortFlag
+		copilotResolvedEffort = origResolvedEffort
 		cli.DryRun = origDryRun
 		copilotCredentialConfigured = origConfigured
 	})
 
 	copilotModelFlag = "gpt-5.5"
+	copilotContextFlag = "default"
+	copilotReasoningEffortFlag = "max"
 	cli.DryRun = true
 	copilotCredentialConfigured = func() bool { return true }
 	cfg := &config.Config{
-		Copilot: config.CopilotConfig{Model: "claude-sonnet-4"},
+		Copilot: config.CopilotConfig{
+			Model:           "claude-sonnet-4",
+			Context:         "long_context",
+			ReasoningEffort: "high",
+		},
 	}
 
 	if err := resolveCopilotPreflight(cfg); err != nil {
@@ -203,6 +235,12 @@ func TestResolveCopilotPreflightModelFlagOverridesConfig(t *testing.T) {
 	if copilotResolvedModel != "gpt-5.5" {
 		t.Fatalf("CLI --model flag should override config model, got %q", copilotResolvedModel)
 	}
+	if copilotResolvedContext != "default" {
+		t.Fatalf("CLI --context flag should override config context, got %q", copilotResolvedContext)
+	}
+	if copilotResolvedEffort != "max" {
+		t.Fatalf("CLI --reasoning-effort flag should override config effort, got %q", copilotResolvedEffort)
+	}
 }
 
 func TestResolveCopilotPreflightAddsRequiredGrantWhenMissing(t *testing.T) {
@@ -250,21 +288,26 @@ func TestResolveCopilotPreflightDryRunDoesNotAddMissingGrant(t *testing.T) {
 }
 
 func TestBuildCopilotCommand(t *testing.T) {
-	origModel, origExperimental, origAutopilot, origAllowAll := copilotResolvedModel, copilotExperimental, copilotAutopilot, copilotAllowAll
+	origModel, origContext, origEffort := copilotResolvedModel, copilotResolvedContext, copilotResolvedEffort
+	origExperimental, origAutopilot, origAllowAll := copilotExperimental, copilotAutopilot, copilotAllowAll
 	t.Cleanup(func() {
 		copilotResolvedModel = origModel
+		copilotResolvedContext = origContext
+		copilotResolvedEffort = origEffort
 		copilotExperimental = origExperimental
 		copilotAutopilot = origAutopilot
 		copilotAllowAll = origAllowAll
 	})
 
 	copilotResolvedModel = "gpt-5.4"
+	copilotResolvedContext = "long_context"
+	copilotResolvedEffort = "high"
 	copilotExperimental = true
 	copilotAutopilot = true
 	copilotAllowAll = true
 
 	got := buildCopilotCommand("fix it", "")
-	want := []string{"copilot", "--no-auto-update", "--model", "gpt-5.4", "--experimental", "--autopilot", "--allow-all", "-p", "fix it"}
+	want := []string{"copilot", "--no-auto-update", "--model", "gpt-5.4", "--context", "long_context", "--reasoning-effort", "high", "--experimental", "--autopilot", "--allow-all", "-p", "fix it"}
 	if !slices.Equal(got, want) {
 		t.Errorf("buildCopilotCommand = %v, want %v", got, want)
 	}
@@ -276,6 +319,17 @@ func TestBuildCopilotCommand(t *testing.T) {
 	if !slices.Contains(got, "-i") || !slices.Contains(got, "hello") {
 		t.Errorf("initial prompt not passed via -i: %v", got)
 	}
+
+	// Companion case: empty context/effort must not produce flags.
+	copilotResolvedContext = ""
+	copilotResolvedEffort = ""
+	got = buildCopilotCommand("fix it", "")
+	if slices.Contains(got, "--context") {
+		t.Errorf("empty context should not produce --context flag: %v", got)
+	}
+	if slices.Contains(got, "--reasoning-effort") {
+		t.Errorf("empty effort should not produce --reasoning-effort flag: %v", got)
+	}
 }
 
 func TestGetCredentialName(t *testing.T) {