view release on metacpan or  search on metacpan

.claude/skills/langertha/SKILL.md  view on Meta::CPAN

### Raid Loop (simplified)

1. Auto-compress history if context threshold exceeded
2. Gather tools from MCP servers + inline tools + self-tools
3. Build conversation: mission + history + new messages
4. Call LLM with tools
5. If tool calls: execute via MCP, add results to conversation, loop
6. If no tool calls: extract final text, persist to history, return result
7. Max iterations safety limit
</raider>

<plugins>
## Plugin System

```perl
package Langertha::Plugin::MyGuardrails;
use Langertha qw( Plugin );

async sub plugin_before_tool_call {
    my ($self, $name, $input) = @_;
    return if $name eq 'dangerous_tool';  # Skip tool
    return ($name, $input);               # Allow tool
}

async sub plugin_after_raid {
    my ($self, $result) = @_;
    return $result;  # Transform result
}

__PACKAGE__->meta->make_immutable;

# Usage
my $raider = Langertha::Raider->new(
    engine  => $engine,
    plugins => ['MyGuardrails', 'Langfuse'],
);
```

### Plugin Hooks (all async sub)

| Hook | Purpose |
|------|---------|
| `plugin_before_raid(@messages)` | Transform input |
| `plugin_build_conversation(@conv)` | Transform assembled conversation |
| `plugin_before_llm_call(@conv, $iter)` | Transform before each LLM call |
| `plugin_after_llm_response($data, $iter)` | Inspect LLM response |
| `plugin_before_tool_call($name, $input)` | Allow/block tool (empty = skip) |
| `plugin_after_tool_call($name, $input, $result)` | Transform tool result |
| `plugin_after_raid($result)` | Transform final result |
</plugins>

<roles>
## Composable Roles

Engines compose feature roles:

| Role | Feature |
|------|---------|
| `Langertha::Role::Chat` | `simple_chat`, `simple_chat_f` |
| `Langertha::Role::Tools` | `chat_with_tools_f` (MCP loop) |
| `Langertha::Role::Streaming` | SSE/NDJSON streaming |
| `Langertha::Role::Embedding` | Vector embeddings |
| `Langertha::Role::Transcription` | Audio-to-text |
| `Langertha::Role::ImageGeneration` | Image generation |
| `Langertha::Role::SystemPrompt` | System prompt management |
| `Langertha::Role::Temperature` | Generation parameters |
| `Langertha::Role::ResponseFormat` | JSON mode / structured output |
| `Langertha::Role::Models` | Model listing |
| `Langertha::Role::Langfuse` | Observability |
| `Langertha::Role::HermesTools` | XML tag tool calling |
| `Langertha::Role::ThinkTag` | Chain-of-thought filtering |
</roles>