Unclear The finding shows the new fields exist but not what setting them actually does.
What
- The
SearchPluginstool got a real description: it now searches your organization's claude.ai plugin catalog proactively, without you needing to name a plugin, whenever a task depends on your team's own process, systems, or data (examples: "ship this to staging", "review this contract against our playbook", "which deals close this week?"). It stays silent if nothing relevant turns up, and won't search again for one-off questions or after you've ignored a previous suggestion. - When results are relevant, Claude calls
SuggestPluginInstallto show an install card; otherwise it relays results in plain text. - The
SuggestPluginInstallcard now records atriggerfield marking whether the suggestion wasproactiveoruser_asked, and Claude is now told to keep working on the task after showing the card instead of stopping. - This
triggervalue is threaded through the system: a related tool's input schema (which dropped a strict enum in favor of looser validation), the plugin listing summary, and theplugin_suggesttelemetry event, so the origin of every suggestion is recorded end to end.
Why
Together these changes let Claude Code recommend org-specific plugins on its own initiative when they'd help with a task, rather than only when explicitly asked, while keeping a record of whether each suggestion was proactive or requested.
Names in the bundlesingle_pagetrigger
The entry above is what we published on the day. These lines were added later, as Anthropic's own pages caught up, and they sit beside the original rather than replacing it.
Added since
A small documentation edit on prompting-claude-fable-5 touched a line naming trigger after this was published.
* **Don't instruct Claude to reproduce its reasoning in the response.** Prompts, skills, or harness instructions that tell the model to echo, transcribe, or explain its internal reasoning as response text can trigger the [`reasoning_extrac…build-with-claude/prompt-engineering/prompting-claude-fable-5 see the edit
Confirmed since
Anthropic's documentation has since written up trigger, on context-editing.
trigger:build-with-claude/context-editing see the edit
Something disagreesSomething we can check disagrees with this entry, or the writer said they could not settle it.
The writer flagged doubt
The finding shows the new fields exist but not what setting them actually does.
Anthropic's documentation agrees
Anthropic's documentation has since written up trigger, on context-editing.