mirror of https://github.com/merlinhu1/codex-game-studio.git synced 2026-08-25 07:54:34 +02:00

Files

T

f3365d2613 feat: close CCGS parity gaps (#12 )

* docs: remove implemented OpenSpec change

Remove the completed eval framework OpenSpec change documents and add a CCGS gap task reference based on the parity reports and upstream CCGS surfaces.

* feat: generate CCGS gap report

Add an OpenSpec-backed implementation for deterministic remaining-gap reporting from the CCGS parity matrix, including tests and regenerated references.

* feat: deepen accessibility role parity

Add OpenSpec-backed CCGS-depth coverage for the accessibility specialist role, including role contract tests, parity regression tests, and regenerated gap reports.

* feat: close direct role depth gaps

Add OpenSpec-backed CCGS-depth contracts for the remaining direct role package gaps, plus parity tests and regenerated remaining-gap reports. Stop before engine sub-specialist, workflow, template, and rule work.

* feat: add engine sub-specialist roles

Add OpenSpec-backed CCGS engine sub-specialist parity, selected-engine role sets, wrong-engine gating, focused engine references, generated custom-agent surfaces, tests, and regenerated parity reports.

* feat: close first workflow parity batch

* feat: close remaining workflow parity gaps

* feat: close final CCGS parity gaps

* chore: remove implemented OpenSpec changes

---------

Co-authored-by: MerlinH <merlinh221@gmail.com>

2026-07-07 14:09:47 +10:00

1.3 KiB

Raw Permalink Blame History

Implementation Agent Protocol

Purpose

Implementation-agent protocol scaffold for code responsibilities, allowed edits, tests, and handoff evidence.

Use this Codex-native template when a game-studio task needs a reviewable implementation-agent-protocol artifact. Keep claims tied to source files, playtest observations, engine output, or explicit user decisions.

Inputs

Project goal, current milestone, and active engine from .codex/studio.json or AGENTS.md.
Relevant design, production, source, asset, or test files named by the task.
Constraints, owners, risks, and required verification evidence.

Outputs

Summary: one paragraph naming the artifact, scope, and decision need.
Source evidence: file paths, commands, observations, or user decisions used.
Artifact body: structured sections appropriate for implementation-agent-protocol with assumptions labeled.
Risks and follow-ups: blockers, unknowns, owner, and due milestone.

Validation

Confirm every required section has concrete content or an explicit Not applicable rationale.
Check paths and commands are repo-relative and safe for the active engine.
Include verification evidence such as tests, build output, playtest notes, review verdict, or handoff.
Record unresolved assumptions separately from confirmed facts.

1.3 KiB Raw Permalink Blame History