feat(gaia): default plan-execution task sub-agents to Sonnet#538
Merged
Conversation
When a plan runs, the cold orchestrator dispatched each task sub-agent with no model pinned, so it inherited the orchestrator session's model (usually Opus). The feature's complexity is already resolved upstream in /gaia-spec and /gaia-plan and their audits, so execution should run on the cheaper model. The /gaia-plan generator now instructs the planner to pin task sub-agents to model: "sonnet" by default in the ORCHESTRATOR.md and KICKOFF.md it writes, with a per-phase Opus escape hatch the planner must justify. Documents the model story in the Task Orchestration and GAIA Plan wiki pages and adds a CHANGELOG entry. Also gitignores Python bytecode from the .gaia/tests observability infra. Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
Contributor
|
code-review-audit skipped: no audit-relevant files changed in the un-audited delta (since the last clean audit); GAIA-Audit commit status stamped on HEAD so the merge gate is satisfied with no local audit run |
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
Contributor
|
code-review-audit skipped: no audit-relevant files changed in the un-audited delta (since the last clean audit); GAIA-Audit commit status stamped on HEAD so the merge gate is satisfied with no local audit run |
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
1 participant
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
What
When a GAIA plan is executed, the cold orchestrator started from
KICKOFF.mddispatched each per-task implementation sub-agent asgeneral-purposewith nomodel:pinned, so it inherited the orchestrator session's model, typically Opus. All the hard reasoning (complexity resolution, contract design, decomposition) is already done upstream in/gaia-spec+ its audit and/gaia-plan+ the decomposition audit, so execution should run on the cheaper model.Change
The
/gaia-plangenerator (.claude/skills/gaia/references/plan.md) now instructs the planner to authorORCHESTRATOR.mdandKICKOFF.mdthat:model: "sonnet"explicitly on dispatch, decoupling execution from the orchestrator's own session model.model: "opus", naming which phase and why. Sonnet is the floor; Opus is a justified exception, never the blanket default.KICKOFF.mdso a cold orchestrator applies it even ifORCHESTRATOR.mddrops from context.The orchestrator itself and the pre-merge
code-review-audit(Opus, pinned via its own frontmatter) are unchanged.Doc-only: the plan artifacts are generated at runtime, so the fix lives in the generator spec plus the
Task OrchestrationandGAIA Planwiki pages and a## [Unreleased]CHANGELOG entry. Also gitignores Python bytecode from the.gaia/testsobservability infra.Verification
.claude/file; no historical/PR/date/UAT/SPEC refs in the edited wiki prose.🤖 Generated with Claude Code