From 787b7fb1aed92c2b701724a7741576053b93cccb Mon Sep 17 00:00:00 2001 From: Claude Sonnet 5 Date: Fri, 3 Jul 2026 23:01:50 +0000 Subject: feat(role): add versioned role configs + escalation ladder + scheduler (Phase 5) Two parts: Part A (fixes a gap from Phase 1): Groq/OpenRouter/OpenAI were documented in docs/api-keys-setup.md as usable once configured, but nothing actually constructed runners for them. internal/cli/cloudrunners.go consolidates anthropic/google/groq/openrouter/openai NativeRunner construction into one table-driven registerCloudRunners() helper, replacing the two hand-written per-provider blocks in serve.go/run.go. Groq/OpenRouter/OpenAI reuse openaicompat (no new adapter code) at SandboxKind: "docker". Part B: the token-husbanding harness's core routing mechanism. - internal/role: RoleConfig/Tier/Rung -- a role's system prompt and a multi-tier (provider, model) escalation ladder, versioned via config_json. - storage: new role_configs table (draft/active/retired, UNIQUE(role, version)) with transactional activate-retires-prior-active semantics; new executions.escalation_rung column. - task.AgentConfig.Role string -- purely additive; every existing task shape (Agent.Role == "") is unaffected, proven by TestPool_Execute_NonRoleTask_Unaffected plus the full pre-existing suite passing unchanged. - executor.Pool.execute(): role-typed tasks with no Agent.Type yet resolve tier 0 of their active ladder (round-robin across multi-candidate tiers, skipping rate-limited providers, falling back to soonest-clearing) before the existing pickAgent/Classifier path runs; SystemPrompt applies to Agent.SystemPromptAppend. Already-resolved role tasks (scheduler resubmits) get their escalation_rung re-derived read-only via findTierIndex. - internal/scheduler: polls role-typed FAILED tasks, retries at the same rung under MaxRetries or escalates to the next tier's first candidate when budget.Accountant.Allow() permits (emitting event.KindEscalated), else leaves the task FAILED with a final:true KindEscalated event. An in-memory per-execution-ID "handled" set keeps the poll loop convergent. Started by `serve` only, config knob [scheduler].poll_interval_seconds. - internal/api: POST/GET /api/roles/{role}/versions, POST /api/roles/{role}/activate -- unauthenticated, matching the existing projects/tasks REST endpoints' auth posture (only chatbot MCP, agent MCP, and WebSocket are api_token-gated in this codebase today). Documented as stored-but-not-yet-enforced (CLAUDE.md Design Debt, matching how task.Priority/RetryConfig are already documented): RoleConfig.Tools/ SandboxKind don't affect dispatch yet; DefaultBudgetUSD is read narrowly as the scheduler's escalation cost estimate, not enforced at initial dispatch; scheduler escalation always targets Candidates[0] (no round-robin, unlike initial-dispatch tier-0 resolution); the scheduler's dedupe is per-process and resets on restart (idempotent, harmless). go build/vet/test -race -count=1 all pass, 21 packages. Co-Authored-By: Claude Sonnet 5 Claude-Session: https://claude.ai/code/session_01V1moSNCJRcP6kykA4tyUSs --- internal/cli/cloudrunners.go | 91 ++++++++++++++++++++++++++++++++++++++++++++ 1 file changed, 91 insertions(+) create mode 100644 internal/cli/cloudrunners.go (limited to 'internal/cli/cloudrunners.go') diff --git a/internal/cli/cloudrunners.go b/internal/cli/cloudrunners.go new file mode 100644 index 0000000..55b3efb --- /dev/null +++ b/internal/cli/cloudrunners.go @@ -0,0 +1,91 @@ +package cli + +import ( + "log/slog" + "net/http" + "time" + + "github.com/thepeterstone/claudomator/internal/config" + "github.com/thepeterstone/claudomator/internal/executor" + "github.com/thepeterstone/claudomator/internal/llm" + "github.com/thepeterstone/claudomator/internal/provider" + "github.com/thepeterstone/claudomator/internal/provider/anthropic" + "github.com/thepeterstone/claudomator/internal/provider/google" + "github.com/thepeterstone/claudomator/internal/provider/openaicompat" +) + +// cloudProviderSpec describes how to construct a provider.Provider for one +// cloud provider entry in cfg.Providers, and which RunnersConfig field gates +// it. anthropic/google use their native wire-format adapters; groq/ +// openrouter/openai are OpenAI-wire-compatible and reuse the Phase 1 +// openaicompat adapter — no dedicated adapter package needed for those three +// (see docs/api-keys-setup.md). +type cloudProviderSpec struct { + name string + enabled func(config.RunnersConfig) bool + build func(pc config.ProviderConfig, timeout time.Duration) provider.Provider +} + +func openaiCompatBuild(pc config.ProviderConfig, timeout time.Duration) provider.Provider { + return openaicompat.New(&llm.Client{ + Endpoint: pc.Endpoint, + Model: pc.DefaultModel, + APIKey: pc.APIKey, + HTTPClient: &http.Client{Timeout: timeout}, + }) +} + +var cloudProviderSpecs = []cloudProviderSpec{ + { + name: "anthropic", + enabled: config.RunnersConfig.AnthropicEnabled, + build: func(pc config.ProviderConfig, timeout time.Duration) provider.Provider { + return anthropic.New(pc.APIKey, pc.Endpoint, timeout) + }, + }, + { + name: "google", + enabled: config.RunnersConfig.GoogleEnabled, + build: func(pc config.ProviderConfig, timeout time.Duration) provider.Provider { + return google.New(pc.APIKey, pc.Endpoint, timeout) + }, + }, + {name: "groq", enabled: config.RunnersConfig.GroqEnabled, build: openaiCompatBuild}, + {name: "openrouter", enabled: config.RunnersConfig.OpenRouterEnabled, build: openaiCompatBuild}, + {name: "openai", enabled: config.RunnersConfig.OpenAIEnabled, build: openaiCompatBuild}, +} + +// registerCloudRunners constructs and adds a NativeRunner (SandboxKind: +// "docker", matching the original anthropic/google-only wiring — cloud +// providers get real container isolation, not the weaker HostSandbox local +// runners use) into runners for every cloud provider in cloudProviderSpecs +// that has a non-empty cfg.Providers[name].APIKey and is enabled via its +// RunnersConfig.Enabled() gate. Called identically from both `serve` +// and `run` — the only difference between them (logDir) is a parameter, and +// neither command does anything provider-specific beyond this. +func registerCloudRunners(runners map[string]executor.Runner, cfg *config.Config, logger *slog.Logger, logDir string) { + for _, spec := range cloudProviderSpecs { + pc, ok := cfg.Providers[spec.name] + if !ok || pc.APIKey == "" || !spec.enabled(cfg.Runners) { + continue + } + timeout := time.Duration(0) + if pc.TimeoutSeconds > 0 { + timeout = time.Duration(pc.TimeoutSeconds) * time.Second + } + runners[spec.name] = &executor.NativeRunner{ + Provider: spec.build(pc, timeout), + Logger: logger, + LogDir: logDir, + DefaultModel: pc.DefaultModel, + // Native cloud providers get real container isolation + // (sandbox.DockerSandbox) rather than HostSandbox's host-side + // path-prefix confinement — see NativeRunner.SandboxKind doc. + SandboxKind: "docker", + SandboxImage: cfg.SandboxImage, + } + if logger != nil { + logger.Info(spec.name+" runner registered", "default_model", pc.DefaultModel, "sandbox", "docker") + } + } +} -- cgit v1.2.3