# `WardrikTriggers.Target`

Connects triggers to your agents. Given to `WardrikTriggers` as `:target`, as `{module, config}`.

Each callback gets `config` as its first argument.

# `ask`

```elixir
@callback ask(
  config :: term(),
  agent_id :: String.t(),
  prompt :: String.t(),
  opts :: [trigger: String.t(), overlap: :skip | :queue, labels: [String.t()]]
) :: :asked | :busy | :no_agent | {:error, term()}
```

Asks `agent_id` `prompt`.

## Options

  * `:trigger` - the trigger's name.
  * `:overlap` - `:skip` to return `:busy` and ask nothing while the agent is busy. `:queue` to
    ask anyway.
  * `:labels` - `["outside"]` when the prompt holds feed entries or a webhook body. `[]`
    otherwise.

Return `:no_agent` if there's no such agent. `{:error, reason}` is logged and reported as
`:failed`.

# `fired`

```elixir
@callback fired(
  config :: term(),
  agent_id :: String.t(),
  trigger :: String.t(),
  WardrikTriggers.Trigger.status()
) :: :ok
```

Called after every run of `trigger`, with how it went.

# `resolve_secret`

```elixir
@callback resolve_secret(config :: term(), agent_id :: String.t(), name :: String.t()) ::
  {:ok, String.t()} | :error
```

Returns the secret `name` held for `agent_id`. A webhook names it as `${NAME}`.

---

*Consult [api-reference.md](api-reference.md) for complete listing*