# PackNotFoundError

`activegraph.packs.load_by_name(name)` searched the `activegraph.packs` entry-point group and found no pack with that name. Either the pack isn't installed, the pack's entry-point group is wrong, or the name is a typo.

This is the third member of the pack-lifecycle cluster, along with [`pack-conflict`](https://docs.activegraph.ai/reference/errors/pack-conflict-error/index.md) (two packs claim the same symbol) and [`pack-version-conflict`](https://docs.activegraph.ai/reference/errors/pack-version-conflict-error/index.md) (same pack, different versions). A developer hitting one of the three might be one step away from hitting another.

## Quick fix

Confirm the pack is installed:

```bash
pip show <pack-distribution-name>
```

List currently-discovered packs:

```python
from activegraph.packs import discover
print([p.name for p in discover()])
```

If the pack is installed but not discovered, its `pyproject.toml` should declare an entry point under the `activegraph.packs` group:

```toml
[project.entry-points."activegraph.packs"]
your_pack = "your_pack_module:pack"
```

The `your_pack_module:pack` form names the Python module where the `Pack` instance lives. Common mistake: pointing the entry point at the module without naming the `pack` attribute.

## How to diagnose

The error message lists what's currently installed:

```text
PackNotFoundError: no installed pack named 'diligence'

What failed:
  activegraph.packs.load_by_name('diligence') searched the
  `activegraph.packs` entry-point group and found no pack with
  that name.
    installed: 'research', 'memory'
```

From code:

```python
try:
    p = load_by_name("diligence")
except PackNotFoundError as e:
    print(e.name)        # 'diligence'
    print(e.installed)   # ('research', 'memory')
```

If `installed` includes the name but `load_by_name` still fails, the entry point is misconfigured — check that the module imports without error and that the named attribute is a `Pack` instance.

## When does this fire

At `activegraph.packs.load_by_name(name)`. Other pack-loading paths (`rt.load_pack(pack_instance, ...)`) take a `Pack` directly and don't go through entry-point discovery, so they don't fire this error.

The CLI's `activegraph quickstart` and similar commands that discover packs by name will surface this if their named pack isn't installed; the error message points at the install command.

## Why the framework refuses to continue

Packs register via Python entry points so the framework can discover them without import-side-effect cost. A missing pack means either the install didn't happen, the entry-point declaration is wrong, or the name is a typo. The runtime refuses to guess — a pack name that doesn't resolve is operator-visible and the recovery is documented.

See [`failure-model`](https://docs.activegraph.ai/concepts/failure-model/index.md) for the broader principle.

## What's related

- [`pack-conflict-error`](https://docs.activegraph.ai/reference/errors/pack-conflict-error/index.md) — fires when two loaded packs declare the same canonical symbol.
- [`pack-version-conflict-error`](https://docs.activegraph.ai/reference/errors/pack-version-conflict-error/index.md) — fires when the runtime already holds a different version of the same pack.
- [Authoring packs](https://docs.activegraph.ai/guides/authoring-packs/index.md) — the canonical pack-format reference, including the `pyproject.toml` entry-point declaration shape.