docs(skills): default PStack to Astra with diverse review models

[BramVR]

PStack uses gpt-6-astra as the default for implementation, explanation, synthesis, and Arena judging, while retaining gpt-5.6-sol for candidate diversity and independent reviews. Every configured role uses high reasoning and standard mode. Fast mode remains forbidden.

• Interrogate: four reviewers using Astra, Sol, Terra, and Luna.
• Arena: judgment-sensitive candidates cycle through Astra, Sol, Terra, and Luna. Generation-bound candidates keep Luna.
• How critique: one architectural critic each on Astra, Sol, and Terra.
• Reflect: Astra judgment and synthesis, Sol divergent review, Luna tooling review.
• Poteto and decision-trail audits: Sol independently reviews Astra work; Terra reviews work performed by Sol.
• Figure-it-out follows Arena's model contract. Poteto subagent instructions honor the assigned role instead of overriding every role to Astra.

Updated the supporting implementation playbooks and changelog. Discovery and procedural verification retain Luna.

Validation: all 89 skills validated; manually reviewed the model contracts, reviewer counts, caller instructions, high reasoning, and no-fast-mode rules. git diff --check passed. Skill-mirror tests and shell syntax checks passed earlier in this PR; this follow-up changes only Markdown instructions.

[Copilot]

🟢 Approval recommended

The changes are consistent, low-risk documentation-only routing updates that align all shown Sol-designated roles to Astra without altering the established Luna/Terra assignments or fast-mode prohibition.

Pull request overview

Updates PStack workflow documentation to route Sol-assigned roles to Astra, aligning model contracts across the skill ecosystem while keeping Luna/Terra assignments and the “no fast mode” rule intact.

Changes:

• Swapped gpt-5.6-sol → gpt-6-astra for implementation/synthesis/judgment roles across PStack workflow skills and Poteto playbooks.
• Updated cross-judge / reviewer model selections where Sol was previously the designated judge/reviewer.
• Added a changelog entry documenting the routing shift.

File summaries

File	Description
skills/why/SKILL.md	Update synthesizer model from Sol to Astra.
skills/show-me-your-work/SKILL.md	Update fallback reviewer model (when Terra cannot review) from Sol to Astra.
skills/reflect/SKILL.md	Update Judgment lens + synthesis model from Sol to Astra.
skills/poteto-mode/SKILL.md	Update Poteto direct-subagent routing for implementation/synthesis/judgment from Sol to Astra.
skills/poteto-mode/references/poteto-agent.md	Update Poteto subagent model from Sol to Astra.
skills/poteto-mode/playbooks/refactoring.md	Update delegated mechanical edit model from Sol to Astra.
skills/poteto-mode/playbooks/perf-issue.md	Update delegated implementation model from Sol to Astra.
skills/poteto-mode/playbooks/orchestrate.md	Update worker/judgment verifier routing from Sol to Astra.
skills/poteto-mode/playbooks/multi-phase-plan.md	Update swarm “gate/perf/audit lane” model from Sol to Astra.
skills/poteto-mode/playbooks/hillclimb.md	Update delegated attempt model from Sol to Astra.
skills/poteto-mode/playbooks/feature.md	Update delegated implementation model from Sol to Astra.
skills/poteto-mode/playbooks/eval.md	Update blinded judge model from Sol to Astra.
skills/poteto-mode/playbooks/bug-fix.md	Update delegated implementation model from Sol to Astra.
skills/interrogate/SKILL.md	Update Reviewer A model from Sol to Astra.
skills/how/SKILL.md	Update explainer/synthesizer/critic model from Sol to Astra.
skills/figure-it-out/SKILL.md	Update architect/arena model reference from Sol to Astra.
skills/arena/SKILL.md	Update runner pool + cross-judge model from Sol to Astra.
CHANGELOG.md	Document the PStack Sol→Astra routing change.

Review details

• Files reviewed: 18/18 changed files
• Comments generated: 0
• Review effort level: Lite

---

💡 Add a code-review agent skill or configure MCP servers for context-aware, tailored reviews. Learn more in the docs.