diff options
| author | Ben Sima <ben@bensima.com> | 2025-12-25 16:34:17 -0500 |
|---|---|---|
| committer | Ben Sima <ben@bensima.com> | 2025-12-25 16:34:17 -0500 |
| commit | 7c9e32f4dee52433911b7e82a72566dc8b5b5708 (patch) | |
| tree | ff6d1cdb87b2383b334d79864d0c9b6e1a7062ff /Omni/Agent/Telegram.hs | |
| parent | 66d2298f29f8e054687acc9e9615ddfa3cdb604a (diff) | |
Omni/Agent/Subagent: add General role and per-spawn customization
- Add SpawnGuardrails type for per-spawn resource limits
- Extend SubagentConfig with toolsOverride, systemPrompt, guardrails
- Add General role with balanced tools (file ops, bash, python, search)
- Update spawn_subagent schema to expose general/custom roles and new params
- Add subagent delegation guidance to Ava's system prompt
🤖 Generated with [Claude Code](https://claude.com/claude-code)
Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>
Diffstat (limited to 'Omni/Agent/Telegram.hs')
| -rw-r--r-- | Omni/Agent/Telegram.hs | 19 |
1 files changed, 19 insertions, 0 deletions
diff --git a/Omni/Agent/Telegram.hs b/Omni/Agent/Telegram.hs index 72cbd6c..2bf7aed 100644 --- a/Omni/Agent/Telegram.hs +++ b/Omni/Agent/Telegram.hs @@ -232,6 +232,25 @@ telegramSystemPromptFallback = "- use this for reminders ('remind me in 2 hours'), follow-ups, or multi-part responses", "- you can list pending messages with 'list_pending_messages' and cancel with 'cancel_message'", "", + "## subagent delegation", + "", + "delegate external actions to subagents rather than executing them directly.", + "this allows you to monitor progress, debug failures, and respawn with adjusted parameters.", + "", + "when to use subagents:", + "- code changes: spawn coder subagent (requires namespace and context)", + "- research tasks: spawn researcher subagent", + "- web scraping: spawn webcrawler subagent", + "- general multi-step tasks: spawn general subagent", + "- custom needs: spawn custom role with specific tools", + "", + "customize each spawn with:", + "- tools: override default tools with specific tool names", + "- system_prompt: additional instructions for this specific task", + "- guardrails: per-spawn limits (max_cost_cents, max_tokens, max_iterations)", + "", + "if a subagent fails, analyze the failure, adjust parameters, and respawn.", + "", "## podcastitlater context", "", "you have access to the PodcastItLater codebase (a product Ben is building) via read_file:", |