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 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01V1moSNCJRcP6kykA4tyUSs

1 files changed, 35 insertions, 10 deletions

diff --git a/internal/storage/db.go b/internal/storage/db.go
index 22a3d7b..a16ad4e 100644
--- a/internal/storage/db.go
+++ b/internal/storage/db.go
@@ -140,6 +140,20 @@ func (s *DB) migrate() error {
 		`CREATE INDEX IF NOT EXISTS idx_events_task_id_seq ON events(task_id, seq)`,
 		`CREATE INDEX IF NOT EXISTS idx_events_ts ON events(ts)`,
 		`CREATE INDEX IF NOT EXISTS idx_events_kind ON events(kind)`,
+		`ALTER TABLE executions ADD COLUMN escalation_rung INTEGER NOT NULL DEFAULT 0`,
+		`CREATE TABLE IF NOT EXISTS role_configs (
+			id TEXT PRIMARY KEY,
+			role TEXT NOT NULL,
+			version INTEGER NOT NULL,
+			status TEXT NOT NULL DEFAULT 'draft',
+			config_json TEXT NOT NULL,
+			created_at DATETIME NOT NULL,
+			activated_at DATETIME,
+			retired_at DATETIME,
+			proposed_by TEXT,
+			UNIQUE(role, version)
+		)`,
+		`CREATE INDEX IF NOT EXISTS idx_role_configs_role_status ON role_configs(role, status)`,
 	}
 	for _, m := range migrations {
 		if _, err := s.db.Exec(m); err != nil {
@@ -522,6 +536,15 @@ type Execution struct {
 	SessionID  string // claude --session-id; persisted for resume
 	SandboxDir string // preserved sandbox path when task is BLOCKED; resume must run here
 	Agent      string // provider that ran this execution (claude/gemini/local); for budget accounting
+	// EscalationRung is the 0-based index into the active role's
+	// EscalationLadder this execution ran at, for role-typed tasks
+	// (task.AgentConfig.Role != ""). 0 (the default) for non-role tasks and
+	// for role-typed executions where the resolved Agent.Type/Model didn't
+	// match any tier in the ladder (e.g. the budget gate rerouted to
+	// "local"). Set once at execution-creation time by
+	// internal/executor.Pool.execute(); internal/scheduler reads it back to
+	// know which tier to retry/escalate from.
+	EscalationRung int
 
 	Changestats *task.Changestats // stored as JSON; nil if not yet recorded
 	Commits     []task.GitCommit // stored as JSON; empty if no commits
@@ -570,10 +593,10 @@ func (s *DB) CreateExecutionAndSetRunning(e *Execution) error {
 		commitsJSON = string(b)
 	}
 	if _, err := tx.Exec(`
-		INSERT INTO executions (id, task_id, start_time, end_time, exit_code, status, stdout_path, stderr_path, artifact_dir, cost_usd, error_msg, session_id, sandbox_dir, changestats_json, commits_json, agent)
-		VALUES (?, ?, ?, ?, ?, ?, ?, ?, ?, ?, ?, ?, ?, NULL, ?, ?)`,
+		INSERT INTO executions (id, task_id, start_time, end_time, exit_code, status, stdout_path, stderr_path, artifact_dir, cost_usd, error_msg, session_id, sandbox_dir, changestats_json, commits_json, agent, escalation_rung)
+		VALUES (?, ?, ?, ?, ?, ?, ?, ?, ?, ?, ?, ?, ?, NULL, ?, ?, ?)`,
 		e.ID, e.TaskID, e.StartTime.UTC(), e.EndTime.UTC(), e.ExitCode, e.Status,
-		e.StdoutPath, e.StderrPath, e.ArtifactDir, e.CostUSD, e.ErrorMsg, e.SessionID, e.SandboxDir, commitsJSON, e.Agent,
+		e.StdoutPath, e.StderrPath, e.ArtifactDir, e.CostUSD, e.ErrorMsg, e.SessionID, e.SandboxDir, commitsJSON, e.Agent, e.EscalationRung,
 	); err != nil {
 		return err
 	}
@@ -633,23 +656,23 @@ func (s *DB) CreateExecution(e *Execution) error {
 		commitsJSON = string(b)
 	}
 	_, err := s.db.Exec(`
-		INSERT INTO executions (id, task_id, start_time, end_time, exit_code, status, stdout_path, stderr_path, artifact_dir, cost_usd, error_msg, session_id, sandbox_dir, changestats_json, commits_json, tokens_in, tokens_out, agent)
-		VALUES (?, ?, ?, ?, ?, ?, ?, ?, ?, ?, ?, ?, ?, ?, ?, ?, ?, ?)`,
+		INSERT INTO executions (id, task_id, start_time, end_time, exit_code, status, stdout_path, stderr_path, artifact_dir, cost_usd, error_msg, session_id, sandbox_dir, changestats_json, commits_json, tokens_in, tokens_out, agent, escalation_rung)
+		VALUES (?, ?, ?, ?, ?, ?, ?, ?, ?, ?, ?, ?, ?, ?, ?, ?, ?, ?, ?)`,
 		e.ID, e.TaskID, e.StartTime.UTC(), e.EndTime.UTC(), e.ExitCode, e.Status,
-		e.StdoutPath, e.StderrPath, e.ArtifactDir, e.CostUSD, e.ErrorMsg, e.SessionID, e.SandboxDir, changestatsJSON, commitsJSON, e.TokensIn, e.TokensOut, e.Agent,
+		e.StdoutPath, e.StderrPath, e.ArtifactDir, e.CostUSD, e.ErrorMsg, e.SessionID, e.SandboxDir, changestatsJSON, commitsJSON, e.TokensIn, e.TokensOut, e.Agent, e.EscalationRung,
 	)
 	return err
 }
 
 // GetExecution retrieves an execution by ID.
 func (s *DB) GetExecution(id string) (*Execution, error) {
-	row := s.db.QueryRow(`SELECT id, task_id, start_time, end_time, exit_code, status, stdout_path, stderr_path, artifact_dir, cost_usd, error_msg, session_id, sandbox_dir, changestats_json, commits_json, tokens_in, tokens_out, agent FROM executions WHERE id = ?`, id)
+	row := s.db.QueryRow(`SELECT id, task_id, start_time, end_time, exit_code, status, stdout_path, stderr_path, artifact_dir, cost_usd, error_msg, session_id, sandbox_dir, changestats_json, commits_json, tokens_in, tokens_out, agent, escalation_rung FROM executions WHERE id = ?`, id)
 	return scanExecution(row)
 }
 
 // ListExecutions returns executions for a task.
 func (s *DB) ListExecutions(taskID string) ([]*Execution, error) {
-	rows, err := s.db.Query(`SELECT id, task_id, start_time, end_time, exit_code, status, stdout_path, stderr_path, artifact_dir, cost_usd, error_msg, session_id, sandbox_dir, changestats_json, commits_json, tokens_in, tokens_out, agent FROM executions WHERE task_id = ? ORDER BY start_time DESC`, taskID)
+	rows, err := s.db.Query(`SELECT id, task_id, start_time, end_time, exit_code, status, stdout_path, stderr_path, artifact_dir, cost_usd, error_msg, session_id, sandbox_dir, changestats_json, commits_json, tokens_in, tokens_out, agent, escalation_rung FROM executions WHERE task_id = ? ORDER BY start_time DESC`, taskID)
 	if err != nil {
 		return nil, err
 	}
@@ -668,7 +691,7 @@ func (s *DB) ListExecutions(taskID string) ([]*Execution, error) {
 
 // GetLatestExecution returns the most recent execution for a task.
 func (s *DB) GetLatestExecution(taskID string) (*Execution, error) {
-	row := s.db.QueryRow(`SELECT id, task_id, start_time, end_time, exit_code, status, stdout_path, stderr_path, artifact_dir, cost_usd, error_msg, session_id, sandbox_dir, changestats_json, commits_json, tokens_in, tokens_out, agent FROM executions WHERE task_id = ? ORDER BY start_time DESC LIMIT 1`, taskID)
+	row := s.db.QueryRow(`SELECT id, task_id, start_time, end_time, exit_code, status, stdout_path, stderr_path, artifact_dir, cost_usd, error_msg, session_id, sandbox_dir, changestats_json, commits_json, tokens_in, tokens_out, agent, escalation_rung FROM executions WHERE task_id = ? ORDER BY start_time DESC LIMIT 1`, taskID)
 	return scanExecution(row)
 }
 
@@ -1123,11 +1146,13 @@ func scanExecution(row scanner) (*Execution, error) {
 	var tokensIn sql.NullInt64
 	var tokensOut sql.NullInt64
 	var agent sql.NullString
+	var escalationRung sql.NullInt64
 	err := row.Scan(&e.ID, &e.TaskID, &e.StartTime, &e.EndTime, &e.ExitCode, &e.Status,
-		&e.StdoutPath, &e.StderrPath, &e.ArtifactDir, &e.CostUSD, &e.ErrorMsg, &sessionID, &sandboxDir, &changestatsJSON, &commitsJSON, &tokensIn, &tokensOut, &agent)
+		&e.StdoutPath, &e.StderrPath, &e.ArtifactDir, &e.CostUSD, &e.ErrorMsg, &sessionID, &sandboxDir, &changestatsJSON, &commitsJSON, &tokensIn, &tokensOut, &agent, &escalationRung)
 	if err != nil {
 		return nil, err
 	}
+	e.EscalationRung = int(escalationRung.Int64)
 	e.SessionID = sessionID.String
 	e.SandboxDir = sandboxDir.String
 	e.TokensIn = tokensIn.Int64