Yeachan-Heo/oh-my-codex · warning · Error

blocked_dependency:${(claim.dependencies ?? []).join(',')}

Error message

blocked_dependency:${(claim.dependencies ?? []).join(',')}

What it means

Thrown when claimTask fails with error 'blocked_dependency'; the message lists the comma-separated task IDs of unresolved dependencies. A task cannot be claimed until its dependencies complete.

Source

Thrown at src/team/runtime.ts:4712

    throw new Error('delegation_only_violation');
  }

  if (governance.plan_approval_required && task.requires_code_change === true) {
    const approved = await isTaskApprovedForExecution(sanitized, taskId, cwd);
    if (!approved) {
      throw new Error('plan_approval_required');
    }
  }
  const config = await readTeamConfig(sanitized, cwd);
  if (!config) throw new Error(`Team ${sanitized} not found`);
  const workerInfo = config.workers.find(w => w.name === workerName);
  if (!workerInfo) throw new Error(`Worker ${workerName} not found in team`);
  const dispatchPolicy = resolveDispatchPolicy(manifest?.policy, config.worker_launch_mode);

  const claim = await claimTask(sanitized, taskId, workerName, task.version ?? 1, cwd);
  if (!claim.ok) {
    if (claim.error === 'blocked_dependency') {
      throw new Error(`blocked_dependency:${(claim.dependencies ?? []).join(',')}`);
    }
    throw new Error(claim.error);
  }

  try {
    // Retry dispatch up to 2 times to handle trust prompts during assignment (fixes #393).
    const approvedExecutionState = await resolvePersistedApprovedTeamExecutionContinuityState(
      sanitized,
      config.leader_cwd ?? cwd,
      config.team_state_root ?? resolveCanonicalTeamStateRoot(config.leader_cwd ?? cwd),
    );
    const persistedUltragoalContext = await readPersistedTeamUltragoalContext(
      sanitized,
      config.leader_cwd ?? cwd,
      config.team_state_root ?? resolveCanonicalTeamStateRoot(config.leader_cwd ?? cwd),
    );
    const approvedContextSection = joinContextSections(
      approvedExecutionState.status === 'valid'

View on GitHub (pinned to 3ad79a8a6f)

Solutions

1. Complete (or skip/cancel) the listed dependency tasks first, then retry assignment
2. If a dependency failed, address its failure or explicitly mark it skipped per the team workflow
3. For independent tasks, remove the incorrect dependency edges from the task definition

Example fix

// before
await assignTask(team, 'T-10', worker, cwd);

// after
// ensure T-8 and T-9 are done first
for (const dep of ['T-8', 'T-9']) await waitForTaskDone(team, dep, cwd);
await assignTask(team, 'T-10', worker, cwd);

const task = await readTask(teamName, taskId, cwd);
const incomplete = (task.dependencies ?? []).filter(d => !isDone(await readTask(teamName, d, cwd)));
if (incomplete.length) throw new Error(`blocked by ${incomplete}`);

try { await assignTask(t, id, w, cwd); } catch (e) { const m = /^blocked_dependency:(.*)$/.exec((e as Error).message); if (m) { await waitForDeps(t, m[1].split(',')); await assignTask(t, id, w, cwd); } else throw e; }

When it happens

Trigger: Assigning a task whose dependency tasks are still pending/blocked/failed; the message encodes claim.dependencies joined by ','.

Common situations: Parallel dispatch of a task graph without respecting dependency order; a dependency failed earlier so dependents stay blocked forever; dependency IDs pointing at deleted tasks.

Related errors

• Team DAG contains a cycle
• Task ${taskId} not found
• claim.error

AI-assisted analysis of Yeachan-Heo/oh-my-codex@3ad79a8a6f (2026-08-27). Data as JSON: /api/errors/c758726452144806. Report an issue: GitHub.