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

Conflicting setup AGENTS merge policy flags: ${source} confl

Error message

Conflicting setup AGENTS merge policy flags: ${source} conflicts with another merge policy selector.

What it means

The `setup` command's AGENTS merge policy can only be selected once. `resolveSetupAgentsMergePolicyArg` throws when two conflicting policy flags are given — i.e. a second flag of a different kind, or two `set`-kind flags with different values (`--merge-agents` vs `--no-merge-agents`).

Source

Thrown at src/cli/index.ts:535

  "team",
  "tmux-hook",
]);

export interface ResolvedCliInvocation {
  command: CliCommand;
  launchArgs: string[];
}

export type SetupMergeAgentsPolicyArg =
  | { kind: "set"; value: boolean }
  | { kind: "clear" }
  | undefined;

export function resolveSetupAgentsMergePolicyArg(args: string[]): SetupMergeAgentsPolicyArg {
  let policy: SetupMergeAgentsPolicyArg;
  const setPolicy = (next: Exclude<SetupMergeAgentsPolicyArg, undefined>, source: string): void => {
    if (policy && (policy.kind !== next.kind || (policy.kind === "set" && next.kind === "set" && policy.value !== next.value))) {
      throw new Error(`Conflicting setup AGENTS merge policy flags: ${source} conflicts with another merge policy selector.`);
    }
    policy = next;
  };
  for (let index = 0; index < args.length; index += 1) {
    const arg = args[index];
    if (arg === "--merge-agents" || arg === "--no-merge-agents" || arg === "--clear-merge-agents-policy") {
      const next = args[index + 1];
      if (next && !next.startsWith("--")) {
        throw new Error(`Setup AGENTS merge policy flags do not accept values: ${arg} ${next}`);
      }
      if (arg === "--merge-agents") {
        setPolicy({ kind: "set", value: true }, arg);
      } else if (arg === "--no-merge-agents") {
        setPolicy({ kind: "set", value: false }, arg);
      } else {
        setPolicy({ kind: "clear" }, arg);
      }
    } else if (

View on GitHub (pinned to 3ad79a8a6f)

Solutions

1. Use exactly one merge policy flag per invocation
2. Pick the policy you actually want: --merge-agents, --no-merge-agents, or --clear-merge-agents-policy
3. Split conflicting policies into separate setup runs if both behaviors are needed over time

Example fix

# before
omx setup --merge-agents --clear-merge-agents-policy

# after
omx setup --merge-agents

const POLICY_FLAGS = ['--merge-agents', '--no-merge-agents', '--clear-merge-agents-policy'];
const used = args.filter(a => POLICY_FLAGS.includes(a));
const uniqueKinds = new Set(used);
if (uniqueKinds.size > 1) {
  console.error('Only one AGENTS merge policy flag is allowed');
  process.exit(1);
}

try {
  await main(args);
} catch (e) {
  if ((e as Error).message.startsWith('Conflicting setup AGENTS merge policy flags')) {
    // strip all but one policy flag and retry
  } else throw e;
}

When it happens

Trigger: Passing more than one of `--merge-agents`, `--no-merge-agents`, `--clear-merge-agents-policy` with differing kinds, e.g. `omx setup --merge-agents --clear-merge-agents-policy` or `omx setup --merge-agents --no-merge-agents`. Repeating the identical flag is fine; conflicting combinations throw.

Common situations: Combining flags copy-pasted from different docs or config templates; shell scripts accumulating options from multiple env sources.

Related errors

• Conflicting setup install mode flags: ${source} selects ${ne
• Conflicting setup MCP mode flags: ${source} selects ${next},
• Conflicting setup Team mode flags: ${source} selects ${next}
• --keep-policy must be one of: score_improvement, pass_only
• Unknown capabilities subcommand: ${parsed.subcommand}

AI-assisted analysis of Yeachan-Heo/oh-my-codex@3ad79a8a6f (2026-08-27). Data as JSON: /api/errors/96c6170323d83385. Report an issue: GitHub.