langchain-ai/deepagents · error · TypeError

hooks trust store must be a JSON object: {path}

Error message

hooks trust store must be a JSON object: {path}

What it means

The hooks trust store file itself must be a JSON object with store metadata. _load_store raises TypeError in strict mode when the parsed file is any other JSON type (array, string, number), and returns an empty store with a warning otherwise. This guards against corrupt or foreign files at the trust-store path.

Source

Thrown at libs/code/deepagents_code/hooks/trust.py:190

        # Decoding happens during the read, so non-UTF-8 stores surface here
        # rather than at `json.loads` below.
        if strict:
            raise
        logger.warning("Could not read hooks trust store %s: %s", path, exc)
        return HooksTrustStore()

    try:
        data: object = json.loads(raw_text)
    except json.JSONDecodeError as exc:
        if strict:
            raise
        logger.warning("Could not parse hooks trust store %s: %s", path, exc)
        return HooksTrustStore()

    if not isinstance(data, dict):
        msg = f"hooks trust store must be a JSON object: {path}"
        if strict:
            raise TypeError(msg)
        logger.warning(msg)
        return HooksTrustStore()

    version = data.get("version")
    if version != _STORE_VERSION:
        msg = f"Unsupported hooks trust store version: {version!r}"
        if strict:
            raise ValueError(msg)
        logger.warning(
            "Ignoring hooks trust store with unsupported version %r", version
        )
        return HooksTrustStore()

    try:
        projects = _parse_projects(data.get("projects"), path=path)
    except TypeError:
        if strict:
            raise

View on GitHub (pinned to a1af029e6e)

Solutions

1. Restore the store to an object shape: {"version": <expected>, "projects": {...}}.
2. Delete the bad store file so the library recreates it, then re-trust the project.
3. Point the trust-store path configuration at the correct file.
4. Call with strict=False if you only need a best-effort check and want a warning plus empty store instead of an exception.

Example fix

// before
store = load_hooks_trust_store(path, strict=True)  # file: ["/repo/a"] -> TypeError

// after
# fix file to:
# {"version": 1, "projects": {"/repo/a": {...}}}
store = load_hooks_trust_store(path, strict=True)

data = json.loads(store_path.read_text())
if not isinstance(data, dict):
    store_path.unlink()  # let library recreate an empty store

def is_store_object(data: object) -> TypeGuard[dict]:
    return isinstance(data, dict)

try:
    store = load_hooks_trust_store(path, strict=True)
except TypeError as exc:
    if "must be a JSON object" in str(exc):
        reset_store_and_retrust(path)

When it happens

Trigger: is_project_hooks_trusted or trust_project_hooks loads a store file whose JSON parses to a non-dict (e.g. a top-level array from an old/foreign format) with strict=True; in non-strict mode this is only a warning.

Common situations: Trust store path pointing at the wrong file (e.g. an unrelated JSON array); a tool writing a different schema; manual edits replacing the object with a list; truncated/corrupt writes (though unparsable JSON is handled earlier).

Related errors

• hooks trust store projects must be an object: {path}
• MCP token file {path} is not a JSON object (found {type(data
• Skill trust store {store_path} is not a JSON object
• {source}: {_display_path(path)}: malformed tools.json: expec
• Failed to parse {SERVER_ENV_PREFIX}{suffix} as JSON: {exc}.

AI-assisted analysis of langchain-ai/deepagents@a1af029e6e (2026-08-29). Data as JSON: /api/errors/f68a1782a2ae7a7d. Report an issue: GitHub.