diff --git a/.polytoken/facets/design.md b/.polytoken/facets/design.md index 6ca9bc5..f232c67 100644 --- a/.polytoken/facets/design.md +++ b/.polytoken/facets/design.md @@ -44,6 +44,6 @@ Use lightweight goal trees only when useful. Stable node IDs and links are enoug When a goal becomes an executable slice, make sure it has goal/non-goals, relevant files/contracts, verified facts, steps or constraints, validation, risks/open decisions, and review expectations. -For small, immediate implementation handoffs, you may use `write_plan` and `handoff_plan` directly, but run the handoff review gate first: at least one `code-reviewer`; for larger plans, the three-reviewer panel with pinned `model_override` values `codex/gpt-5.5(medium)`, `anthropic/claude-opus-4-8(medium)`, and `zai/glm-5.2(max)`. Fix or rebut every Critical/High finding, re-run until clean, and stop for operator input if a pinned model does not resolve. +For small, immediate implementation handoffs, you may use `write_plan` and `handoff_plan` directly, but run the handoff review gate first: at least one `code-reviewer`; for larger plans, the three-reviewer panel with pinned `model_override` values `codex/gpt-5.5(medium)`, `anthropic/claude-opus-4-8(medium)`, and `zai/glm-5.2(max)`. Fix or rebut every Critical/High finding, re-run until clean, and stop for operator input if a pinned model does not resolve. Ensure you tell the reviewer the exact absolute path to the plan file to read it. For larger implementation planning, switch to `plan` so the dedicated planning facet can produce and review the execution handoff. diff --git a/.polytoken/facets/plan.md b/.polytoken/facets/plan.md index 0aabee4..ea9627d 100644 --- a/.polytoken/facets/plan.md +++ b/.polytoken/facets/plan.md @@ -26,6 +26,7 @@ Before `handoff_plan`, run an independent review gate: - For small/simple handoffs, run at least one `plan-reviewer` subagent against the handoff plan. - For larger plans, run three reviewers in parallel using `code-reviewer` with `model_override` values `codex/gpt-5.5(medium)`, `anthropic/claude-opus-4-8(medium)`, and `zai/glm-5.2(max)`. - If a pinned `model_override` does not resolve at runtime, stop and ask the operator; do not silently substitute a weaker model. +- ensure you tell the reviewer the exact absolute path to the plan file to read it. - Fix or rebut every Critical/High finding with evidence. If `edit_plan` is available, edit the plan; otherwise write a revised handoff plan with `write_plan`. - Re-run the same reviewer set while any reviewer returns any Critical or High finding. - Only call `handoff_plan` when all reviewers report zero Critical and zero High findings, or when the operator explicitly accepts the risk. diff --git a/.polytoken/project_vars.yaml b/.polytoken/project_vars.yaml index bcde12f..9770750 100644 --- a/.polytoken/project_vars.yaml +++ b/.polytoken/project_vars.yaml @@ -41,6 +41,6 @@ orual_workflow: review: gate_severities: [Critical, High] reviewer_models: - - codex/gpt-5.5 - - codex/gpt-5.4-mini - - zai/glm-5.2 + - codex/gpt-5.5(medium) + - codex/gpt-5.4-mini(xhigh) + - zai/glm-5.2(max) diff --git a/.polytoken/skills/jira-solo-tasker-execute/SKILL.md b/.polytoken/skills/jira-solo-tasker-execute/SKILL.md index 5de4eda..c7ada03 100644 --- a/.polytoken/skills/jira-solo-tasker-execute/SKILL.md +++ b/.polytoken/skills/jira-solo-tasker-execute/SKILL.md @@ -100,8 +100,8 @@ Use the narrowest sufficient checks first, then broader checks if the change war Run the configured reviewer set. Default panel: -- `code-reviewer` with `model_override` `codex/gpt-5.5(high)` -- `code-reviewer` with `model_override` `anthropic/claude-opus-4-8(medium)` +- `code-reviewer` with `model_override` `codex/gpt-5.5(medium)` +- `code-reviewer` with `model_override` `codex/gpt-5.4-mini(xhigh)` - `code-reviewer` with `model_override` `zai/glm-5.2(max)` Ask reviewers for Critical/High/Medium/Low findings, with evidence and suggested fixes. diff --git a/.polytoken/skills/jira-solo-tasker-plan/SKILL.md b/.polytoken/skills/jira-solo-tasker-plan/SKILL.md index f16c30e..d9dbb30 100644 --- a/.polytoken/skills/jira-solo-tasker-plan/SKILL.md +++ b/.polytoken/skills/jira-solo-tasker-plan/SKILL.md @@ -125,7 +125,7 @@ The exact parent directory for sibling workspaces is project policy. If not obvi Before `handoff_plan`, run three `plan-reviewer` subagents in parallel with the configured/pinned models: - `codex/gpt-5.5(medium)` -- `anthropic/claude-opus-4-8(medium)` +- `codex/gpt-5.4-mini(xhigh)` - `zai/glm-5.2(max)` Fix or rebut every plan finding with evidence. Re-run the same reviewer set while any reviewer returns any Critical or High finding. diff --git a/.polytoken/subagents/code-reviewer.md b/.polytoken/subagents/code-reviewer.md index 698a79e..108afb1 100644 --- a/.polytoken/subagents/code-reviewer.md +++ b/.polytoken/subagents/code-reviewer.md @@ -3,7 +3,7 @@ name: code-reviewer description: Adversarial reviewer for implementation plans, code changes, tests, and shipping quality. Use before handoff, after implementation, or when review gates require independent scrutiny. polytoken: model: default_model:full - tools: [file_read, grep, glob, shell_exec, skill] + tools: [file_read, grep, glob, shell_exec, skill, mcp__atlassian] undeferred_tools: [file_read, grep, glob] skills_allow: ["reviewing-code-for-shipping-quality", "polytoken:investigating-a-codebase"] ---