Repository navigation
feat(pstack): upgrade models and require three architect candidates - #21
Merged
Merged
Conversation
Contributor
There was a problem hiding this comment.
Copilot review overview
🟡 Changes recommended
Arena’s “proceed with N-1 on dropout” guidance still conflicts with the new Architect requirement to have one complete candidate per configured runner before synthesis.
Get a fresh assessment by requesting another Copilot review.
Review effort: Lite
Findings: 1
What changed in this PR
Update pstack skill defaults and docs to move active Sol/Luna roles to GPT-6, and tighten Architect to require a full three-runner panel before synthesis.
Changes:
- Upgrade defaults across routed skills/playbooks from
gpt-5.6-{sol,luna}togpt-6-{sol,luna}; update the plan validator lane string accordingly. - Expand Arena default runner panel and standardize Claude invocation via
claude -p(with a Claude Code version requirement for Opus). - Update Architect guidance to require three distinct design candidates, one per configured runner, before synthesis (but see review comments re: dropout behavior).
| File | Description |
|---|---|
| skills/why/SKILL.md | Update investigator/synthesizer default models to GPT-6. |
| skills/swarm/SKILL.md | Update default swarm worker model to GPT-6 Luna. |
| skills/setup-pstack/SKILL.md | Align setup template with new per-role defaults and Claude CLI guidance. |
| skills/reflect/SKILL.md | Update tooling reflection default model to GPT-6 Sol. |
| skills/poteto-mode/SKILL.md | Update global spawn defaults for code model to GPT-6 Sol. |
| skills/poteto-mode/scripts/check-plan.mjs | Update required “Ten lanes on …” string to GPT-6 Luna. |
| skills/poteto-mode/references/poteto-agent.md | Update poteto-agent implementation model to GPT-6 Sol. |
| skills/poteto-mode/references/models.md | Update canonical per-role model defaults and Arena/Architect panels. |
| skills/poteto-mode/playbooks/refactoring.md | Update delegate model default to GPT-6 Sol. |
| skills/poteto-mode/playbooks/perf-issue.md | Update delegate model default to GPT-6 Sol. |
| skills/poteto-mode/playbooks/multi-phase-plan.md | Update live verification lane model references to GPT-6 Luna. |
| skills/poteto-mode/playbooks/hillclimb.md | Update delegate model default to GPT-6 Sol. |
| skills/poteto-mode/playbooks/feature.md | Update delegate model default to GPT-6 Sol. |
| skills/poteto-mode/playbooks/bug-fix.md | Update delegate model default to GPT-6 Sol. |
| skills/no-comments/SKILL.md | Update Comment Sicko default model to GPT-6 Luna. |
| skills/interrogate/SKILL.md | Update Reviewer B default model to GPT-6 Sol. |
| skills/how/SKILL.md | Update explorer/explainer default models to GPT-6. |
| skills/codex-first/SKILL.md | Update pinned Codex exec default model to GPT-6 Sol. |
| skills/arena/SKILL.md | Update default runner panel + Claude invocation guidance. |
| skills/architect/SKILL.md | Require three Architect design candidates (one per runner) before synthesis. |
| CHANGELOG.md | Document model upgrades and the Architect 3-candidate requirement. |
💡 Add a code-review agent skill or configure MCP servers for context-aware, tailored reviews. Learn more in the docs.
Comment on lines
+32
to
+34
| Use `architect runners` from `${CODEX_HOME:-$HOME/.codex}/skills/poteto-mode/references/models.md` when present. Otherwise default to one each on `gpt-6-sol` at `high` reasoning, `gpt-6-astra` at `low` reasoning, `claude-fable-5-1` at `high` reasoning. Pass this panel to Arena as the explicit runner list. | ||
|
|
||
| Design it twice. Require at least two structurally distinct candidates before synthesis, even when the first looks sufficient. This is the **exhaust-the-design-space** principle skill made concrete. Whole-shape alternatives, not point fixes inside one shape. | ||
| Design it three times. Require three structurally distinct candidates before synthesis, one from each configured Architect runner, even when the first looks sufficient. This is the **exhaust-the-design-space** principle skill made concrete. Whole-shape alternatives, not point fixes inside one shape. |
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
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.
Architect could proceed with only two sketches despite its configured three-model panel. Require three distinct candidates, one per Architect runner, before synthesis.
Upgrade active Sol/Luna roles to GPT-6 at high reasoning and align skill defaults, setup templates, and the live-plan validator. Arena uses Astra medium, GPT-6 Sol high, GPT-5.6 Sol high, Opus 5.5 high, and Fable 5.1 high. Claude candidates use
claude -p; Opus requires CLI 2.1.280+.Validation: 106 skills validated; skill-mirror tests, shell syntax, plan-validator marker checks, downloader smoke checks, and diff checks passed. Real Architect dry run completed Sol high, Astra low, and Fable 5.1 high candidates before Astra medium cross-judging and synthesis. Fable response metadata verified; separate Opus CLI probe passed. No queue implementation was performed.