Slipstream/auto_review
1
0
Fork 0
Code Issues 2 Pull requests Projects Releases 27 Packages Activity Actions

chore(opencode): configure per-agent model routing and plan-advisor #245

Merged
jwilger merged 1 commit from chore/opencode-model-routing into main 2026-05-18 09:40:16 -07:00
Conversation 5 Commits 1 Files changed 3 +69
jwilger commented 2026-05-18 09:05:52 -07:00
Owner
Copy link

Summary

  • Set global default and per-agent model routing in opencode.json.
  • Route build and RGR reviewer agents to openai/gpt-5.3-codex-spark.
  • Route plan, architecture, security, and plan-advisor to openai/gpt-5.5 with higher reasoning effort.
  • Add a read-only .opencode/agents/plan-advisor.md subagent profile.
  • Update .opencode/agents/build.md to escalate ambiguous or cross-cutting RGR decisions to @plan-advisor.

Why

This change enables higher reasoning capability where it matters while keeping routine verification/build tasks on a faster model profile, improving cost and latency characteristics without reducing review quality.

Verification

  • just fmt
  • opencode debug config
## Summary - Set global default and per-agent model routing in `opencode.json`. - Route build and RGR reviewer agents to `openai/gpt-5.3-codex-spark`. - Route `plan`, `architecture`, `security`, and `plan-advisor` to `openai/gpt-5.5` with higher reasoning effort. - Add a read-only `.opencode/agents/plan-advisor.md` subagent profile. - Update `.opencode/agents/build.md` to escalate ambiguous or cross-cutting RGR decisions to `@plan-advisor`. ## Why This change enables higher reasoning capability where it matters while keeping routine verification/build tasks on a faster model profile, improving cost and latency characteristics without reducing review quality. ## Verification - `just fmt` - `opencode debug config`
chore(opencode): add agent model routing and plan advisor
All checks were successful
CI / Format check (pull_request) Successful in 5s
Details
CI / Clippy (pull_request) Successful in 41s
Details
CI / Dependency policy (pull_request) Successful in 12s
Details
CI / Test (pull_request) Successful in 54s
Details
CI / Build (pull_request) Successful in 33s
Details
CI / Request auto_review semantic review (pull_request) Successful in 1s
Details
CI / Build PR artifacts (no token) (pull_request) Successful in 2s
Details
auto_review auto_review: no findings
24b26d9a0f
auto-review requested changes 2026-05-18 09:07:31 -07:00
Dismissed
auto-review left a comment
Copy link

This PR configures per-agent model routing in opencode.json, introducing a new plan-advisor subagent for high-reasoning tasks. The changes appear safe to merge, with no apparent issues in the configuration.

Walkthrough

  • .opencode/agents/build.md:

    • Added a directive to route ambiguous or cross-cutting decisions to @plan-advisor for high-reasoning analysis.

  • .opencode/agents/plan-advisor.md:

    • Introduced a new subagent configuration for plan-advisor, designed for high-reasoning tasks with specific permissions and constraints.

  • opencode.json:

    • Configured global and per-agent model routing, specifying different models for various agents, including openai/gpt-5.5 for high-reasoning tasks.

Pre-merge checks

  • PR metadata quality: failed
  • Rationale: The PR body contains a validation section that is not relevant to the change and does not explain why the changes are needed. It appears to be a checklist or notes rather than a meaningful description.
  • Offending text: ## Validation
  • just fmt
  • opencode debug config
This PR configures per-agent model routing in `opencode.json`, introducing a new `plan-advisor` subagent for high-reasoning tasks. The changes appear safe to merge, with no apparent issues in the configuration. ## Walkthrough - **.opencode/agents/build.md**: - Added a directive to route ambiguous or cross-cutting decisions to `@plan-advisor` for high-reasoning analysis. - **.opencode/agents/plan-advisor.md**: - Introduced a new subagent configuration for `plan-advisor`, designed for high-reasoning tasks with specific permissions and constraints. - **opencode.json**: - Configured global and per-agent model routing, specifying different models for various agents, including `openai/gpt-5.5` for high-reasoning tasks. ## Pre-merge checks - PR metadata quality: failed - Rationale: The PR body contains a validation section that is not relevant to the change and does not explain why the changes are needed. It appears to be a checklist or notes rather than a meaningful description. - Offending text: ## Validation - just fmt - opencode debug config
jwilger commented 2026-05-18 09:38:52 -07:00
Author
Owner
Copy link

Updated the PR body to match metadata quality criteria: added explicit Summary and Why sections with rationale, while keeping validation commands in a proper section so the pre-merge checker should no longer flag this as a failure. Please rerun PR checks/re-review.

Updated the PR body to match metadata quality criteria: added explicit `Summary` and `Why` sections with rationale, while keeping validation commands in a proper section so the pre-merge checker should no longer flag this as a failure. Please rerun PR checks/re-review.
jwilger commented 2026-05-18 09:39:38 -07:00
Author
Owner
Copy link

@auto-review re-review

@auto-review re-review
auto-review commented 2026-05-18 09:39:38 -07:00
Owner
Copy link

Queued a forced review at 24b26d9a0f. This intentionally bypasses CI gating; watch the commit-status badge for progress.

Queued a forced review at 24b26d9a0fece1fa6810804ca90c96f3d2839ae8. This intentionally bypasses CI gating; watch the commit-status badge for progress.
auto-review approved these changes 2026-05-18 09:39:44 -07:00
auto-review left a comment
Copy link

The PR configures per-agent model routing in opencode.json, optimizing for reasoning capability where needed. It introduces a new plan-advisor subagent profile and updates the build agent to escalate certain decisions. The changes appear safe to merge.

Walkthrough

  • .opencode/agents/build.md:
    • Updated to route ambiguous or cross-cutting decisions to @plan-advisor for high-reasoning analysis.
  • .opencode/agents/plan-advisor.md:
    • Introduced a new subagent profile for complex planning and ambiguous tradeoffs, using openai/gpt-5.5.
  • opencode.json:
    • Configured global and per-agent model routing, assigning openai/gpt-5.5 to agents requiring higher reasoning effort.
The PR configures per-agent model routing in `opencode.json`, optimizing for reasoning capability where needed. It introduces a new `plan-advisor` subagent profile and updates the `build` agent to escalate certain decisions. The changes appear safe to merge. ## Walkthrough - **.opencode/agents/build.md**: - Updated to route ambiguous or cross-cutting decisions to `@plan-advisor` for high-reasoning analysis. - **.opencode/agents/plan-advisor.md**: - Introduced a new subagent profile for complex planning and ambiguous tradeoffs, using `openai/gpt-5.5`. - **opencode.json**: - Configured global and per-agent model routing, assigning `openai/gpt-5.5` to agents requiring higher reasoning effort.
jwilger deleted branch chore/opencode-model-routing 2026-05-18 09:40:17 -07:00
Sign in to join this conversation.
Reviewers
No reviewers
auto-review
Labels
Clear labels   
bug

   
discussion

   
enhancement
  
bug
Something is not working

  
duplicate
This issue or pull request already exists

  
enhancement
New feature

  
help wanted
Need some help

  
invalid
Something is wrong

  
question
More information is needed

  
wontfix
This won't be fixed

No labels
bug
discussion
enhancement
bug
duplicate
enhancement
help wanted
invalid
question
wontfix
Milestone
Clear milestone
No items
No milestone
Projects
Clear projects
No items
No project
Assignees
Clear assignees
No assignees
2 participants
Notifications
Due date
The due date is invalid or out of range. Please use the format "yyyy-mm-dd".

No due date set.

Dependencies

No dependencies set.

Reference
Slipstream/auto_review!245
No description provided.