can1357/oh-my-pi · error · ValueError

model must be 'provider/model' (e.g. anthropic/claude-sonnet

Error message

model must be 'provider/model' (e.g. anthropic/claude-sonnet-4-6)

What it means

OmpLocalAgent.run requires self.model_name in 'provider/model' form because it splits on '/' to derive the CLI --provider and --model arguments. If model_name is missing (None/empty) or contains no slash, run raises this ValueError before launching the agent.

Source

Thrown at packages/metaharness/agent/omp_local.py:545

        env: dict[str, str] = {}
        for key in _PROVIDER_KEYS.get(provider, []):
            value = os.environ.get(key)
            if value:
                env[key] = value
        return env

    # ---------------------------------------------------------------------- run

    @with_prompt_template
    @override
    async def run(
        self,
        instruction: str,
        environment: BaseEnvironment,
        context: AgentContext,
    ) -> None:
        if not self.model_name or "/" not in self.model_name:
            raise ValueError(
                "model must be 'provider/model' (e.g. anthropic/claude-sonnet-4-6)"
            )
        provider, model = self.model_name.split("/", 1)

        if self._binary:
            parts = [shlex.quote(self._cli)]
        else:
            parts = [shlex.quote(self._bun), shlex.quote(self._cli)]
        parts += [
            "--print",
            "--mode json",
            f"--provider {shlex.quote(provider)}",
            f"--model {shlex.quote(model)}",
            "--no-session",
        ]
        if self._auto_approve:
            parts.append("--auto-approve")
        if self._thinking:

View on GitHub (pinned to 9690622007)

Solutions

1. Set the model to a fully qualified 'provider/model' string, e.g. anthropic/claude-sonnet-4-6.
2. Pass the model at construction time of OmpLocalAgent (or via its config/env) so model_name is populated.
3. Validate the model string at startup rather than inside run to fail earlier.
4. If your model id itself contains no slash, prefix the provider: e.g. 'openai/gpt-5'.

Example fix

# before
agent = OmpLocalAgent(model_name="claude-sonnet-4-6")
# after
agent = OmpLocalAgent(model_name="anthropic/claude-sonnet-4-6")

def ensure_provider_model(name):
    if not name or "/" not in name:
        raise ValueError(f"model must be 'provider/model', got {name!r}")
    provider, model = name.split("/", 1)
    return provider, model

def is_provider_model(name) -> bool:
    return isinstance(name, str) and "/" in name and all(name.split("/", 1))

try:
    await agent.run(instruction, env, ctx)
except ValueError as e:
    if "provider/model" in str(e):
        agent.model_name = f"anthropic/{agent.model_name}"  # or reconfigure
        return await agent.run(instruction, env, ctx)
    raise

When it happens

Trigger: Calling run() with the agent constructed without a model, or with a bare model id like 'claude-sonnet-4-6' lacking the provider prefix.

Common situations: Config file/env omitted the model; user specified only the model name assuming a default provider; programmatic use passed a raw model id without the anthropic/ or openai/ prefix.

Understand the failure class

Background: "Invalid ... format", "must be in format X", "does not look like a ..." — invalid argument format errors across CLI tools and libraries — this error's family across 17 libraries.

Related errors

• OMP_AUTH_BROKER_ACCOUNT_POOL_FILE contains an empty provider
• OMP_AUTH_BROKER_ACCOUNT_POOL_FILE contains a provider id wit
• ${name} path does not exist: ${trimmed}
• Anthropic thinking budget requires max_tokens greater than $
• cachedContent must not be blank

AI-assisted analysis of can1357/oh-my-pi@9690622007 (2026-08-31). Data as JSON: /api/errors/56d7096509c6aab9. Report an issue: GitHub.