merge: v2.6.3 — GPT-5.6 codex defaults (Sol/Terra/Luna) + stale adapter pin refresh

Author: copeus

.claude-plugin/marketplace.json

@@ -9,7 +9,7 @@
     {
       "name": "superpowers-v",
       "description": "Compound V for Superpowers: triple parallel pre-flight (code archaeology + domain-expert + Context7 library validator), disjoint partitioning, manifest-driven multi-backend dispatch (Claude + Codex + Antigravity + Cursor), git-diff scope enforcement, crash-resumable runs, adaptive tier-based routing, epic mode, V-memory local-first semantic+lexical recall over docs/superpowers (opt-in pure-python embeddings + a deterministic recall→action bridge), and batched parallel dispatch (Opus default, narrow Sonnet exception)",
-      "version": "2.6.2",
+      "version": "2.6.3",
       "source": "./",
       "author": {
         "name": "Oleg",

.claude-plugin/plugin.json

@@ -1,7 +1,7 @@
 {
   "name": "superpowers-v",
   "description": "Compound V for Superpowers: triple parallel pre-flight (code archaeology + domain-expert advisor + library/doc validator via Context7), disjoint file partitioning, manifest-driven multi-backend dispatch (Claude + headless Codex + Antigravity + Cursor workers), git-diff scope enforcement, crash-resumable runs, adaptive tier-based routing, epic mode for multi-feature builds, V-memory local-first semantic+lexical recall over docs/superpowers (opt-in pure-python embeddings + a deterministic recall\u2192action bridge), and batched parallel dispatch (Opus default, Sonnet for narrow junior-task carve-out), plus /v:onboard — a project-onboarding command that builds a citation-verified knowledge base + AGENTS.md/CLAUDE.md bridge behind a human gate. Auto-intercepts brainstorming \u2192 writing-plans \u2192 execution transitions.",
-  "version": "2.6.2",
+  "version": "2.6.3",
   "author": {
     "name": "Oleg",
     "email": "copeus@gmail.com"

CHANGELOG.md

@@ -4,6 +4,14 @@ All notable changes to **superpowers-v (Compound V)** are documented here.
 
 The format is based on [Keep a Changelog](https://keepachangelog.com/en/1.1.0/), and this project uses semantic versioning.
 
+## [2.6.3] — 2026-07-10
+
+### Changed — Codex defaults bumped to the GPT-5.6 family (Sol/Terra/Luna)
+
+- **`deep`→`gpt-5.6-sol`, `standard`→`gpt-5.6-terra`, `light`→`gpt-5.6-luna`** (was `gpt-5.5`/`gpt-5.5`/`gpt-5.3-codex-spark`) — a real per-tier differentiation where `deep`/`standard` previously shared the same model. Live-verified all three on `codex-cli 0.144.1` (`PROBE_OK`). **`gpt-5.6-sol` requires codex-cli >= 0.143.0** — confirmed broken with a clear 400 `"requires a newer version of Codex"` on 0.142.5, working on 0.144.1; an under-floor client fails loud (not silently — the failure-policy retries once then halts cleanly). `gpt-5.6-terra`/`gpt-5.6-luna` work on older clients too (verified back to 0.142.5). `compound-v-codex-review.sh`'s cross-model-review default follows the deep tier (`gpt-5.6-sol`, "Codex on their max").
+- **Two independently-stale adapter pins refreshed during the audit:** `adapter-codex.md`'s verified-against pin (`0.130.0` → `0.144.1`); `adapter-cursor.md`'s verified-against pin (`2025.09.12` → `2026.06.26`) **and** a now-**false** claim — "cursor-agent has no models list command" — corrected: it does now (`cursor-agent models`, a live 187-entry catalog verified). **Grok is not present in that live catalog** for this account (press coverage says available, likely region/plan-gated — not documented since unconfirmed hands-on). No auto-discovery was added for Cursor's catalog (would be over-engineering — it spans unrelated vendor families with no shared naming convention, unlike Antigravity's single-family Gemini catalog `/v:models` already ranks); curated + user-overridable stays the flow, now pointing at the real command for manual discovery.
+- Every doc stating the codex model map as **current fact** updated for consistency (`v-init.md` seed, `v-models.md` roster/table/example, `routing-policy.md` map/resolve-example, `execution-manifest.md` tier table/config example). Illustrative "(e.g. `gpt-5.5`)" mentions explaining the resolution *mechanism* (never hardcode a model — the resolver handles it) and the dated `routing-lessons.md` historical entry were deliberately left untouched. **Codex cross-model verification caught one real miss** — `compound-v-resolve-model.py`'s own source comment still said "cursor-agent has no `models` list command" (I'd audited the `.md` docs for this false claim but missed the `.py` comment) — fixed; a second pass confirmed every codex tier mapping, the review-script default, and every version pin consistent, with no remaining stale `gpt-5.5`/`gpt-5.3-codex-spark` current-default claims.
+
 ## [2.6.2] — 2026-07-06
 
 ### Fixed — `.claude/compound-v.json` no longer commits machine-local capability

commands/v-init.md

@@ -106,12 +106,13 @@ cursor-agent status </dev/null 2>&1 | head -3   # or: [ -n "$CURSOR_API_KEY" ]
 ```
 
 - Installed **and** authenticated → Cursor is **usable**; record it and add it to `backends`.
-  The pinned headless invocation holds (verified, cursor-agent 2025.09.12):
+  The pinned headless invocation holds (verified, cursor-agent 2026.06.26):
   `cd "$WT" && cursor-agent -p -f --output-format json [--model <M>] "<prompt>" </dev/null`
   (`.result` → summary, `.session_id` → resume). Default model is **`auto`** — VERIFIED that a
   Cursor **Free** plan can *only* use Auto (named models like `sonnet-4` error). On a **paid**
-  plan, set named per-tier ids via [`/v:models`](v-models.md) / config (no `models` list
-  command, so no auto-discovery). Note the plan when you record it.
+  plan, run `cursor-agent models` to see the live catalog, then set named per-tier ids via
+  [`/v:models`](v-models.md) / config (manual — Compound V doesn't auto-rank cursor's multi-vendor
+  catalog). Note the plan when you record it.
 - Installed but **not** authenticated → record it as **present but unauthenticated**; treat as
   unavailable and tell the user to run `cursor-agent login` (or set `CURSOR_API_KEY`).
 
@@ -278,13 +279,13 @@ was, in those two fields).
   "models": {
     "balanced": {
       "claude":      { "deep": "opus",                  "standard": "opus",                  "light": "sonnet" },
-      "codex":       { "deep": "gpt-5.5",               "standard": "gpt-5.5",               "light": "gpt-5.3-codex-spark" },
+      "codex":       { "deep": "gpt-5.6-sol",            "standard": "gpt-5.6-terra",          "light": "gpt-5.6-luna" },
       "antigravity": { "deep": "Gemini 3.1 Pro (High)", "standard": "Gemini 3.1 Pro (Low)", "light": "Gemini 3.5 Flash (Low)" },
       "cursor":      { "deep": "auto",                  "standard": "auto",                  "light": "auto" }
     },
     "cost-aware": {
       "claude":      { "deep": "opus",                  "standard": "sonnet",                "light": "sonnet" },
-      "codex":       { "deep": "gpt-5.5",               "standard": "gpt-5.5",               "light": "gpt-5.3-codex-spark" },
+      "codex":       { "deep": "gpt-5.6-sol",            "standard": "gpt-5.6-terra",          "light": "gpt-5.6-luna" },
       "antigravity": { "deep": "Gemini 3.1 Pro (High)", "standard": "Gemini 3.1 Pro (Low)", "light": "Gemini 3.5 Flash (Low)" },
       "cursor":      { "deep": "auto",                  "standard": "auto",                  "light": "auto" }
     }

commands/v-models.md

@@ -37,13 +37,13 @@ everywhere else `standard` Claude is `opus`:
 "models": {
   "balanced": {
     "claude":      { "deep": "opus",                  "standard": "opus",                  "light": "sonnet" },
-    "codex":       { "deep": "gpt-5.5",               "standard": "gpt-5.5",               "light": "gpt-5.3-codex-spark" },
+    "codex":       { "deep": "gpt-5.6-sol",            "standard": "gpt-5.6-terra",          "light": "gpt-5.6-luna" },
     "antigravity": { "deep": "Gemini 3.1 Pro (High)", "standard": "Gemini 3.1 Pro (Low)", "light": "Gemini 3.5 Flash (Low)" },
     "cursor":      { "deep": "auto",                  "standard": "auto",                  "light": "auto" }
   },
   "cost-aware": {
     "claude":      { "deep": "opus",                  "standard": "sonnet",                "light": "sonnet" },
-    "codex":       { "deep": "gpt-5.5",               "standard": "gpt-5.5",               "light": "gpt-5.3-codex-spark" },
+    "codex":       { "deep": "gpt-5.6-sol",            "standard": "gpt-5.6-terra",          "light": "gpt-5.6-luna" },
     "antigravity": { "deep": "Gemini 3.1 Pro (High)", "standard": "Gemini 3.1 Pro (Low)", "light": "Gemini 3.5 Flash (Low)" },
     "cursor":      { "deep": "auto",                  "standard": "auto",                  "light": "auto" }
   }
@@ -81,8 +81,9 @@ the user override any entry by hand (a model the curated list doesn't know about
 still valid — codex accepts whatever model string you pass to `codex exec --model`).
 Present this curated starting roster:
 
-- `gpt-5.5` — strongest; suggested for `deep` and `standard`
-- `gpt-5.3-codex-spark` — fast/cheap; suggested for `light`
+- `gpt-5.6-sol` — strongest; suggested for `deep` (requires codex-cli >= 0.143.0)
+- `gpt-5.6-terra` — balanced; suggested for `standard`
+- `gpt-5.6-luna` — fast/cheap; suggested for `light`
 
 Confirm codex is even usable first (so you don't write a map the project can't run):
 
@@ -128,11 +129,16 @@ command -v agy >/dev/null \
   The script never invents names — it only ranks the catalog `agy models` actually
   printed, so anything you show came from the live CLI.
 
-### 1d. cursor — Auto by default (no list command; plan-gated)
+### 1d. cursor — Auto by default (manual list command; plan-gated)
 
-cursor-agent has **no `models` list command**, and named models are **plan-gated**: a Cursor
-**Free** plan can only use **`auto`** (passing a named model errors with *"Named models
-unavailable"* — verified live). So the default map is `auto` for every tier:
+cursor-agent (2026.06.26+) has a `models` list command (`cursor-agent models`) for **manual**
+discovery — a paid-plan user can run it to see the live, real catalog and pick a named override.
+Compound V does not auto-rank/auto-discover it (unlike antigravity's single-family Gemini
+catalog): cursor's catalog spans many unrelated vendor families (GPT/Claude/Gemini/…) with no
+shared naming/effort convention, so ranking it well would need its own bespoke logic — curated +
+user-overridable stays the flow. Named models are also **plan-gated**: a Cursor **Free** plan can
+only use **`auto`** (passing a named model errors with *"Named models unavailable"* — verified
+live). So the default map is `auto` for every tier:
 
 ```bash
 command -v cursor-agent && cursor-agent status </dev/null >/dev/null 2>&1 && echo "cursor usable (auth ok)" || echo "cursor unavailable/unauthed"
@@ -156,7 +162,7 @@ fast/cheap option → `light`). Example shape:
 | Backend | Available now | deep | standard | light |
 |---|---|---|---|---|
 | claude | opus, sonnet | opus | opus | sonnet |
-| codex | gpt-5.5, gpt-5.3-codex-spark | gpt-5.5 | gpt-5.5 | gpt-5.3-codex-spark |
+| codex | gpt-5.6-sol, gpt-5.6-terra, gpt-5.6-luna | gpt-5.6-sol | gpt-5.6-terra | gpt-5.6-luna |
 | antigravity | *(from `agy models </dev/null`)* | Gemini 3.1 Pro (High) | Gemini 3.1 Pro (Low) | Gemini 3.5 Flash (Low) |
 
 Then **let the user assign** each tier per backend — accept the suggestion as-is, or
@@ -197,13 +203,13 @@ Resulting shape (only `models` is this command's responsibility) — write the
   "models": {
     "balanced": {
       "claude":      { "deep": "opus",    "standard": "opus",    "light": "sonnet" },
-      "codex":       { "deep": "gpt-5.5", "standard": "gpt-5.5", "light": "gpt-5.3-codex-spark" },
+      "codex":       { "deep": "gpt-5.6-sol", "standard": "gpt-5.6-terra", "light": "gpt-5.6-luna" },
       "antigravity": { "deep": "…",       "standard": "…",       "light": "…" },
       "cursor":      { "deep": "auto",    "standard": "auto",    "light": "auto" }
     },
     "cost-aware": {
       "claude":      { "deep": "opus",    "standard": "sonnet",  "light": "sonnet" },
-      "codex":       { "deep": "gpt-5.5", "standard": "gpt-5.5", "light": "gpt-5.3-codex-spark" },
+      "codex":       { "deep": "gpt-5.6-sol", "standard": "gpt-5.6-terra", "light": "gpt-5.6-luna" },
       "antigravity": { "deep": "…",       "standard": "…",       "light": "…" },
       "cursor":      { "deep": "auto",    "standard": "auto",    "light": "auto" }
     }

scripts/compound-v-codex-review.sh

@@ -25,16 +25,18 @@
 #     [--schema <abs>] [--context-file <abs>] ... \
 #     [--timeout-sec <n>]
 #
-# Defaults: model gpt-5.5, effort high (the "Codex on their max" the design calls for),
-# schema = <plugin>/schemas/plan-review.schema.json, resolved relative to THIS script —
-# the reviewed repo has no reason to carry the plugin's schema.
+# Defaults: model gpt-5.6-sol, effort high (the "Codex on their max" the design calls for;
+# requires codex-cli >= 0.143.0 -- an older client fails loud with a clear "requires a newer
+# version of Codex" error, not silently), schema = <plugin>/schemas/plan-review.schema.json,
+# resolved relative to THIS script — the reviewed repo has no reason to carry the plugin's
+# schema.
 #
 # Exit: 0 when findings JSON was produced (even verdict=reject — that is reported IN the
 # JSON). Non-zero only on a usage / environment fault.
 
 set -euo pipefail
 
-DEFAULT_MODEL="gpt-5.5"
+DEFAULT_MODEL="gpt-5.6-sol"
 DEFAULT_EFFORT="high"
 DEFAULT_TIMEOUT_SEC=600
 

scripts/compound-v-resolve-model.py

@@ -23,7 +23,7 @@
                                     light→low when --effort omitted)
 
 Output: a single JSON object on stdout, e.g.
-  {"backend": "codex", "tier": "deep", "model": "gpt-5.5", "effort": "high"}
+  {"backend": "codex", "tier": "deep", "model": "gpt-5.6-sol", "effort": "high"}
 
 Exit non-zero if a tier cannot be resolved for a backend (and no
 --explicit-model was given).
@@ -54,7 +54,11 @@
 # codex/antigravity/cursor are identical across stances. NEVER 'haiku' anywhere.
 _CLAUDE_DEFAULT = {"deep": "opus", "standard": "opus", "light": "sonnet"}
 _CLAUDE_COST_AWARE = {"deep": "opus", "standard": "sonnet", "light": "sonnet"}
-_CODEX = {"deep": "gpt-5.5", "standard": "gpt-5.5", "light": "gpt-5.3-codex-spark"}
+# GPT-5.6 family (Sol/Terra/Luna), verified live 2026-07-10: all three confirmed working on
+# codex-cli 0.144.1. gpt-5.6-sol specifically requires codex-cli >= 0.143.0 (confirmed: broken
+# with a clear 400 "requires a newer version of Codex" on 0.142.5, works on 0.144.1) -- an
+# under-floor client fails LOUD (not silent; the failure-policy retries once then halts cleanly).
+_CODEX = {"deep": "gpt-5.6-sol", "standard": "gpt-5.6-terra", "light": "gpt-5.6-luna"}
 # Antigravity (agy): FALLBACK default; the live catalog is discoverable headlessly
 # (`agy models </dev/null`), and /v:models/+/v:init pipe it through
 # compound-v-discover-models.py to OVERRIDE this map in .claude/compound-v.json. Names
@@ -65,8 +69,9 @@
                 "light": "Gemini 3.5 Flash (Low)"}
 # Cursor (cursor-agent): "auto" is the SAFE DEFAULT for every tier — a FREE plan can ONLY
 # use Auto (named models error: "Named models unavailable"). Paid plans override per-tier
-# via /v:models (cursor-agent has no `models` list command). Lower-trust tier (no kernel
-# sandbox; headless -f required).
+# via /v:models — `cursor-agent models` lists the live catalog for manual discovery (not
+# auto-ranked: it spans unrelated vendor families with no shared naming convention).
+# Lower-trust tier (no kernel sandbox; headless -f required).
 _CURSOR = {"deep": "auto", "standard": "auto", "light": "auto"}