Configuration reference

A one-page index of every setting that changes what the model reads on a turn — where it’s set, what it defaults to, and what it changes. Use this page to find the right lever quickly, then follow the link to the full reference for the syntax and examples. ABL settings take effect in new conversations once saved, or once published and deployed for production channels. Project settings take effect on the next turn unless noted otherwise.

Knowledge-base and voice levers are covered in full on Knowledge bases and knowledge tools and Voice agents — this table includes only their headline entries.

Setting	Where	Default	Changes
`BEHAVIOR: role:`	Agent ABL	Not set (allow-list mode)	Which values reach `## Current Context`, and which prompt template is used — see How a turn is built
`GOAL`, `PERSONA`, `LIMITATIONS`, `{{refs}}`	Agent ABL	—	The identity section; references also make values visible in `## Current Context`
`SYSTEM_PROMPT:`	Agent ABL	Catalog template	Replaces the template; runtime sections are still appended
`MEMORY: SESSION:`	Agent ABL	—	`## Session Memory` content, and what the model may write
`MEMORY: persistent:`, `remember:`, `recall:`	Agent ABL	—	Values loaded into the session; `recall: INSTRUCTION` lines in the prompt
`HANDOFF … CONTEXT:` (`pass`, `summary`, `history`, `memory_grants`)	Agent ABL	`history: full`	The receiving agent’s history, context, and granted memory
`DELEGATE … INPUT` / `PURPOSE`	Agent ABL	—	The only context a delegate gets
`CONVERSATION:`	Agent ABL	—	`## Conversation Behavior` lines; listening and barge-in behavior on voice
`EXECUTION.inline_gather`	Agent ABL	`false` (on in the voice profile)	`## Information Collection` instead of the `GATHER` list
`EXECUTION.compaction.*`	Agent ABL	See Context size, compaction, and budgets	Tool results, earlier turns, and `## Current Context` budgets
`EXECUTION.models.summarization`	Agent ABL	Project tier	The model used for tool-result summaries
`EXECUTION.voice.instructions`	Agent ABL	—	`## Voice Delivery Instructions` and voice formatting rules
`EXECUTION.guardrail_builtins`	Agent ABL	`false`	Whether a guardrail check can read `content`, `session_state`, and `tool_results`
Tool `MODEL_CONTEXT` (`FIELDS` / `TEMPLATE`)	Tool ABL	—	What the model sees of a tool result, now and later
Response `MODEL_CONTEXT`	Hook `RESPOND`, `COMPLETE`, `FLOW` step	—	What the history records for a rich response
Tool `compaction:` (`essential_fields`, `summary_timeout_ms`, `max_search_tokens`)	Tool ABL	—	Per-tool compaction, and the knowledge-base token budget
Tool `store_result`	Tool ABL	`true`; `false` once the tool has `on_result: SET`	Whether `last_<tool>_result` is kept
Tool param `default_source`, `visibility_model`, `visibility_trace`	Tool ABL	Visible	Hides a runtime-supplied value from the model and from traces
KB `search_instructions`, `kb_name`, `description`	Tool ABL / KB settings	—	How and when the model searches
Index Top K, similarity threshold, keyword weight	Studio KB index settings	10, 0.2, 0.5	How many chunks come back
Citations	Studio KB citation settings	On by default, per index	Citation instructions added to the prompt
`GUARDRAILS:`	Agent ABL	—	Per-agent checks — see Guardrail checkpoints
Guardrail policies, providers, fail mode	Studio Govern → Guardrails	Fail open	Project and tenant checks; `## Current Policy`
Project compaction	Project runtime config (API or import)	Platform defaults	Project-wide budgets; resolved once per conversation
Voice latency profile and overrides	Studio Settings → Runtime Config	Auto-seeded for voice	Tighter budgets, fast model tier, reasoning off, inline gather on
Filler	Project runtime config	Profile defaults	Voice filler behavior and the status-update instruction
PII redaction (`pii_redaction`)	Project runtime config	On	Tokens in place of personal data in the prompt and in traces. Built-in detectors start off on new projects.
Multi-intent	Project runtime config	—	Multi-request routing and task sections
Cross-channel recall	Project omnichannel settings	Per project	`## Previous Conversations (Other Channels)`: up to 30 messages of 500 characters each

Fixed by the platform, not configurable: the order of prompt sections, the Customer Output Contract, the tool-history marker and its 6-round window, the realtime voice handoff history-replay limits, and the absolute limits on ## Current Context (40,000 characters per value, 100,000 for the whole block).

Related articles:

How a turn is built — what each of these settings ultimately controls.
Context size, compaction, and budgets — the compaction defaults in full.
Agent declaration — GOAL/PERSONA/SYSTEM_PROMPT and the EXECUTION: block.
Multi-agent and supervisor — HANDOFF, DELEGATE, and ESCALATE.