@ai-dev-methodologies/rlp-desk 0.9.3 → 0.10.0

package/README.md

@@ -244,8 +244,8 @@ When all US pass individually, the final ALL verify runs **sequentially per-US**
 | `--verifier-model MODEL` | sonnet | per-US verification model (lighter) |
 | `--final-verifier-model MODEL` | opus | final ALL verification model (stricter) |
 | `--consensus off\|all\|final-only` | off | Cross-engine consensus scope |
-| `--consensus-model MODEL` | gpt-5.4:medium | per-US cross-verifier (lighter) |
-| `--final-consensus-model MODEL` | gpt-5.4:high | final cross-verifier (stricter) |
+| `--consensus-model MODEL` | gpt-5.5:medium | per-US cross-verifier (lighter) |
+| `--final-consensus-model MODEL` | gpt-5.5:high | final cross-verifier (stricter) |
 | `--verify-mode per-us\|batch` | per-us | per-us: verify each US → final ALL |
 | `--cb-threshold N` | 6 | Consecutive failures → BLOCKED |
 | `--max-iter N` | 100 | Max iterations → TIMEOUT |
@@ -267,7 +267,7 @@ When `--consensus` is enabled, a second cross-engine verifier runs alongside eac
 After `brainstorm`, `init` detects your environment and presents run command presets:
 
 - **Codex detected (GPT Pro / spark)** → recommends cross-engine mode (`--worker-model spark:high --consensus final-only`)
-- **Codex detected (large PRD, AC > 15)** → offers gpt-5.4 preset (`--worker-model gpt-5.4:high --consensus final-only`)
+- **Codex detected (large PRD, AC > 15)** → offers gpt-5.5 preset (`--worker-model gpt-5.5:high --consensus final-only`)
 - **Claude-only** → defaults to `--debug` with haiku worker and opus final verifier
 - **Basic** → minimal flags for quick iteration
 
@@ -363,7 +363,7 @@ npm install -g @openai/codex
 /rlp-desk run calculator --worker-model spark:high
 
 # Customize model and reasoning effort
-/rlp-desk run calculator --worker-model gpt-5.4:high
+/rlp-desk run calculator --worker-model gpt-5.5:high
 
 # Cross-engine: codex worker, claude verifier (recommended)
 /rlp-desk run calculator --worker-model spark:high --consensus final-only --debug
@@ -406,7 +406,7 @@ By default, Worker and Verifier stop and ask for human input when they encounter
 **`--autonomous`** enables fully unattended campaigns:
 
 ```bash
-/rlp-desk run my-feature --mode tmux --worker-model gpt-5.4:medium --autonomous --debug
+/rlp-desk run my-feature --mode tmux --worker-model gpt-5.5:medium --autonomous --debug
 ```
 
 ### How it works

package/docs/protocol-reference.md

@@ -438,7 +438,7 @@ The `run` command accepts engine flags to control which CLI executes Worker and
 |------|---------|-------------|
 | `--worker-engine claude\|codex` | `claude` | Engine for Worker |
 | `--verifier-engine claude\|codex` | `claude` | Engine for Verifier |
-| `--codex-model MODEL` | `gpt-5.4` | Model passed to the `codex` CLI (when engine=codex) |
+| `--codex-model MODEL` | `gpt-5.5` | Model passed to the `codex` CLI (when engine=codex) |
 | `--codex-reasoning low\|medium\|high` | `high` | Reasoning effort for the `codex` CLI |
 
 **Claude engine** (default): uses `Agent()` in agent mode, `claude -p` with `--dangerously-skip-permissions` in tmux mode.

package/docs/superpowers/plans/2026-04-24-gpt-5-5-default.md

@@ -0,0 +1,517 @@
+# Default Codex Model Migration (gpt-5.4 → gpt-5.5) Implementation Plan
+
+> **For agentic workers:** REQUIRED SUB-SKILL: Use superpowers:subagent-driven-development (recommended) or superpowers:executing-plans to implement this plan task-by-task. Steps use checkbox (`- [ ]`) syntax for tracking.
+
+**Goal:** Migrate rlp-desk's default Codex model references and env-var defaults from `gpt-5.4` to `gpt-5.5` across 6 source files, without adding remap logic or altering reasoning-effort tiers. Ship as v0.10.0.
+
+**Architecture:** Purely textual replacement in configuration sites (shell defaults, Node defaults, upgrade-chain keys, doc examples, upgrade-table markdown) plus a `version` bump. No runtime logic is added. User-supplied model strings are unchanged — Codex CLI already accepts any passthrough name.
+
+**Tech Stack:** zsh, Node.js (ESM), markdown, npm (package.json). No new dependencies.
+
+**Branch:** `feature/gpt-5-5-default` (already created at plan time).
+
+**Spec:** `docs/superpowers/specs/2026-04-24-gpt-5-5-default.md`.
+
+---
+
+## File Map
+
+| File | Role after this plan |
+|---|---|
+| `src/scripts/run_ralph_desk.zsh` | Shell entrypoint. Env var defaults now quote `gpt-5.5` in four places and in two CLI fallback sites. |
+| `src/node/run.mjs` | Node CLI defaults. `consensusModel` and `finalConsensusModel` default to `gpt-5.5:medium` / `gpt-5.5:high`. |
+| `src/node/runner/campaign-main-loop.mjs` | Upgrade chain keys use `gpt-5.5` namespace. Claude-engine and spark rows unchanged. |
+| `src/model-upgrade-table.md` | "Non-Pro" section retitled to `gpt-5.5`; table cells use `gpt-5.5:*`. |
+| `src/commands/rlp-desk.md` | Recommendation table, worker-model guidance, batch warnings, examples, and help text reference `gpt-5.5`. |
+| `package.json` | `version` bumped from `0.9.3` to `0.10.0`. |
+| `docs/superpowers/specs/2026-04-24-gpt-5-5-default.md` | Already written; intentionally contains `gpt-5.4` as historical context. Excluded from regression sweep. |
+
+Self-verification gate (per `CLAUDE.md`) is a required workflow step before commit; it does not produce tracked files but is captured as Task 8.
+
+---
+
+## Task 1: Update shell defaults in `run_ralph_desk.zsh`
+
+**Files:**
+- Modify: `src/scripts/run_ralph_desk.zsh:33,35,104,106,114-115,2762,2766`
+
+- [ ] **Step 1: Replace comment defaults on lines 33 and 35**
+
+Before (lines 33 and 35):
+
+```zsh
+#   WORKER_CODEX_MODEL            - codex model for Worker (default: gpt-5.4)
+...
+#   VERIFIER_CODEX_MODEL          - codex model for Verifier (default: gpt-5.4)
+```
+
+After:
+
+```zsh
+#   WORKER_CODEX_MODEL            - codex model for Worker (default: gpt-5.5)
+...
+#   VERIFIER_CODEX_MODEL          - codex model for Verifier (default: gpt-5.5)
+```
+
+Edit via the Edit tool, matching the full line so the change is unambiguous.
+
+- [ ] **Step 2: Replace env var defaults on lines 104 and 106**
+
+Before:
+
+```zsh
+WORKER_CODEX_MODEL="${WORKER_CODEX_MODEL:-gpt-5.4}"
+WORKER_CODEX_REASONING="${WORKER_CODEX_REASONING:-high}"   # low|medium|high
+VERIFIER_CODEX_MODEL="${VERIFIER_CODEX_MODEL:-gpt-5.4}"
+```
+
+After:
+
+```zsh
+WORKER_CODEX_MODEL="${WORKER_CODEX_MODEL:-gpt-5.5}"
+WORKER_CODEX_REASONING="${WORKER_CODEX_REASONING:-high}"   # low|medium|high
+VERIFIER_CODEX_MODEL="${VERIFIER_CODEX_MODEL:-gpt-5.5}"
+```
+
+- [ ] **Step 3: Replace consensus defaults on lines 114 and 115**
+
+Before:
+
+```zsh
+CONSENSUS_MODEL="${CONSENSUS_MODEL:-gpt-5.4:medium}"       # per-US cross-verifier (lighter)
+FINAL_CONSENSUS_MODEL="${FINAL_CONSENSUS_MODEL:-gpt-5.4:high}"  # final cross-verifier (stricter)
+```
+
+After:
+
+```zsh
+CONSENSUS_MODEL="${CONSENSUS_MODEL:-gpt-5.5:medium}"       # per-US cross-verifier (lighter)
+FINAL_CONSENSUS_MODEL="${FINAL_CONSENSUS_MODEL:-gpt-5.5:high}"  # final cross-verifier (stricter)
+```
+
+- [ ] **Step 4: Replace CLI fallback defaults on lines 2762 and 2766**
+
+Before:
+
+```zsh
+    --consensus-model)
+      (( _cli_i++ ))
+      CONSENSUS_MODEL="${@[$_cli_i]:-gpt-5.4:medium}"
+      ;;
+    --final-consensus-model)
+      (( _cli_i++ ))
+      FINAL_CONSENSUS_MODEL="${@[$_cli_i]:-gpt-5.4:high}"
+      ;;
+```
+
+After:
+
+```zsh
+    --consensus-model)
+      (( _cli_i++ ))
+      CONSENSUS_MODEL="${@[$_cli_i]:-gpt-5.5:medium}"
+      ;;
+    --final-consensus-model)
+      (( _cli_i++ ))
+      FINAL_CONSENSUS_MODEL="${@[$_cli_i]:-gpt-5.5:high}"
+      ;;
+```
+
+- [ ] **Step 5: Verify zero residual `gpt-5.4` in file**
+
+Run:
+
+```bash
+grep -n "gpt-5\.4" src/scripts/run_ralph_desk.zsh
+```
+
+Expected: no output. Any hit means the previous steps missed a line.
+
+- [ ] **Step 6: Shell syntax sanity check**
+
+Run:
+
+```bash
+zsh -n src/scripts/run_ralph_desk.zsh
+```
+
+Expected: no output (exit 0). Parse errors would be shown to stderr.
+
+---
+
+## Task 2: Update Node defaults in `run.mjs`
+
+**Files:**
+- Modify: `src/node/run.mjs:14-15`
+
+- [ ] **Step 1: Replace default model strings in `RUN_DEFAULTS`**
+
+Before (lines 14-15):
+
+```js
+  consensusModel: 'gpt-5.4:medium',
+  finalConsensusModel: 'gpt-5.4:high',
+```
+
+After:
+
+```js
+  consensusModel: 'gpt-5.5:medium',
+  finalConsensusModel: 'gpt-5.5:high',
+```
+
+- [ ] **Step 2: Verify zero residual `gpt-5.4` in file**
+
+Run:
+
+```bash
+grep -n "gpt-5\.4" src/node/run.mjs
+```
+
+Expected: no output.
+
+- [ ] **Step 3: Parse check**
+
+Run:
+
+```bash
+node --check src/node/run.mjs
+```
+
+Expected: no output (exit 0). Syntax errors would print to stderr.
+
+---
+
+## Task 3: Update upgrade chain in `campaign-main-loop.mjs`
+
+**Files:**
+- Modify: `src/node/runner/campaign-main-loop.mjs:29-31`
+
+- [ ] **Step 1: Replace the non-spark rows of `MODEL_UPGRADES`**
+
+Before (lines 28-35):
+
+```js
+const MODEL_UPGRADES = {
+  'gpt-5.4:medium': 'gpt-5.4:high',
+  'gpt-5.4:high': 'gpt-5.4:xhigh',
+  'gpt-5.4:xhigh': 'BLOCKED',
+  'gpt-5.3-codex-spark:medium': 'gpt-5.3-codex-spark:high',
+  'gpt-5.3-codex-spark:high': 'gpt-5.3-codex-spark:xhigh',
+  'gpt-5.3-codex-spark:xhigh': 'BLOCKED',
+};
+```
+
+After:
+
+```js
+const MODEL_UPGRADES = {
+  'gpt-5.5:medium': 'gpt-5.5:high',
+  'gpt-5.5:high': 'gpt-5.5:xhigh',
+  'gpt-5.5:xhigh': 'BLOCKED',
+  'gpt-5.3-codex-spark:medium': 'gpt-5.3-codex-spark:high',
+  'gpt-5.3-codex-spark:high': 'gpt-5.3-codex-spark:xhigh',
+  'gpt-5.3-codex-spark:xhigh': 'BLOCKED',
+};
+```
+
+- [ ] **Step 2: Verify zero residual `gpt-5.4` in file**
+
+Run:
+
+```bash
+grep -n "gpt-5\.4" src/node/runner/campaign-main-loop.mjs
+```
+
+Expected: no output.
+
+- [ ] **Step 3: Parse check**
+
+Run:
+
+```bash
+node --check src/node/runner/campaign-main-loop.mjs
+```
+
+Expected: no output.
+
+---
+
+## Task 4: Update upgrade table in `model-upgrade-table.md`
+
+**Files:**
+- Modify: `src/model-upgrade-table.md:21-28`
+
+- [ ] **Step 1: Rename the Non-Pro section and rewrite the table body**
+
+Before (lines 21-28):
+
+```markdown
+## Non-Pro (gpt-5.4)
+
+| Complexity | 1-2 | 3-4 | 5-6 | 7+ |
+|------------|-----|-----|-----|-----|
+| LOW | gpt-5.4:low | gpt-5.4:medium | gpt-5.4:high | BLOCKED |
+| MEDIUM | gpt-5.4:medium | gpt-5.4:high | gpt-5.4:xhigh | BLOCKED |
+| HIGH | gpt-5.4:high | gpt-5.4:xhigh | gpt-5.4:xhigh | BLOCKED |
+| CRITICAL | gpt-5.4:xhigh | gpt-5.4:xhigh | gpt-5.4:xhigh | BLOCKED |
+```
+
+After:
+
+```markdown
+## Non-Pro (gpt-5.5)
+
+| Complexity | 1-2 | 3-4 | 5-6 | 7+ |
+|------------|-----|-----|-----|-----|
+| LOW | gpt-5.5:low | gpt-5.5:medium | gpt-5.5:high | BLOCKED |
+| MEDIUM | gpt-5.5:medium | gpt-5.5:high | gpt-5.5:xhigh | BLOCKED |
+| HIGH | gpt-5.5:high | gpt-5.5:xhigh | gpt-5.5:xhigh | BLOCKED |
+| CRITICAL | gpt-5.5:xhigh | gpt-5.5:xhigh | gpt-5.5:xhigh | BLOCKED |
+```
+
+The GPT Pro and Claude-only sections remain unchanged.
+
+- [ ] **Step 2: Verify zero residual `gpt-5.4` in file**
+
+Run:
+
+```bash
+grep -n "gpt-5\.4" src/model-upgrade-table.md
+```
+
+Expected: no output.
+
+---
+
+## Task 5: Update user-facing documentation in `rlp-desk.md`
+
+**Files:**
+- Modify: `src/commands/rlp-desk.md` (14 known occurrences across ~6 blocks: lines 78-81, 84, 94-95, 167, 173, 180, 185-186, 243, 251-252, 295-296, 476-477, 735, 740-741)
+
+Because the same token (`gpt-5.4`) appears with both `:medium` and `:high` suffixes in a tabular document, use a global replace rather than editing each occurrence individually.
+
+- [ ] **Step 1: Perform global replace of `gpt-5.4` → `gpt-5.5`**
+
+Use the Edit tool with `replace_all: true` on the single token `gpt-5.4`. This is safe because:
+- No reasoning effort label matches `gpt-5.4` as a substring.
+- No URL or external reference in the file uses `gpt-5.4`.
+- The spec file that intentionally mentions `gpt-5.4` lives in `docs/superpowers/specs/`, not `src/commands/`.
+
+Edit parameters:
+
+```
+old_string: "gpt-5.4"
+new_string: "gpt-5.5"
+replace_all: true
+```
+
+- [ ] **Step 2: Verify zero residual `gpt-5.4` in file**
+
+Run:
+
+```bash
+grep -n "gpt-5\.4" src/commands/rlp-desk.md
+```
+
+Expected: no output.
+
+- [ ] **Step 3: Spot-check a few critical lines**
+
+Run:
+
+```bash
+grep -n "gpt-5\.5" src/commands/rlp-desk.md | head -20
+```
+
+Expected: at least 14 lines referencing `gpt-5.5`, including the recommendation table body (two rows with `gpt-5.5:medium`, two with `gpt-5.5:high`), the worker model selection bullet, batch warnings, env var echo lines, example commands, and the help text.
+
+---
+
+## Task 6: Bump version to 0.10.0
+
+**Files:**
+- Modify: `package.json:3`
+
+- [ ] **Step 1: Edit the `version` field**
+
+Before:
+
+```json
+  "version": "0.9.3",
+```
+
+After:
+
+```json
+  "version": "0.10.0",
+```
+
+- [ ] **Step 2: Verify package.json is still valid JSON**
+
+Run:
+
+```bash
+node -e 'const pkg = require("./package.json"); console.log(pkg.version)'
+```
+
+Expected stdout: `0.10.0`.
+
+---
+
+## Task 7: Repo-wide regression sweep
+
+**Files:** none modified in this task — it is a verification gate.
+
+- [ ] **Step 1: Confirm zero `gpt-5.4` hits under `src/`**
+
+Run:
+
+```bash
+grep -rn "gpt-5\.4" src/
+```
+
+Expected: no output.
+
+- [ ] **Step 2: Confirm zero `gpt-5.4` hits under `docs/` except the spec file**
+
+Run:
+
+```bash
+grep -rn "gpt-5\.4" docs/ --exclude-dir=specs
+```
+
+Expected: no output.
+
+And:
+
+```bash
+grep -rln "gpt-5\.4" docs/superpowers/specs/
+```
+
+Expected: exactly one file — `docs/superpowers/specs/2026-04-24-gpt-5-5-default.md`.
+
+- [ ] **Step 3: Confirm `gpt-5.5` references now exist**
+
+Run:
+
+```bash
+grep -rc "gpt-5\.5" src/ | grep -v ':0$'
+```
+
+Expected: at least 5 files each with a non-zero count, covering `run_ralph_desk.zsh`, `run.mjs`, `campaign-main-loop.mjs`, `model-upgrade-table.md`, and `commands/rlp-desk.md`.
+
+If any step here fails, stop and repair the offending file before continuing. Do not proceed to Task 8 until the sweep is clean.
+
+---
+
+## Task 8: Self-verification gate (CLAUDE.md §Self-Verification Gate)
+
+**Files:** none modified directly. This task produces three verifier verdict records that must all PASS before Task 9.
+
+This task is mandatory because Tasks 1 and 5 modified `src/scripts/run_ralph_desk.zsh` and `src/commands/rlp-desk.md`. Per `CLAUDE.md`, all three scenarios must run with real Worker (`execution_steps`) → Verifier (5 categories with `reasoning`) → PASS. Any FAIL blocks commit; fix and re-run all three.
+
+- [ ] **Step 1: Scenario LOW — default env propagation (L1+L3)**
+
+Describe:
+- **Worker task:** From the repo root, source `src/scripts/run_ralph_desk.zsh` (or simulate by evaluating only the env-var default block, lines 104-115 and 2760-2768) and print `WORKER_CODEX_MODEL`, `VERIFIER_CODEX_MODEL`, `CONSENSUS_MODEL`, `FINAL_CONSENSUS_MODEL`. Execution steps must include the exact commands run and their stdout.
+- **Verifier categories:** (1) Correctness — all four variables equal the expected `gpt-5.5*` strings. (2) Completeness — no variable is empty or inherits an unexpected default. (3) Regression — the script still parses under `zsh -n`. (4) Docs alignment — the printed value matches what `rlp-desk.md` now documents. (5) Reasoning — one paragraph explaining why the result is convincing.
+- **PASS criteria:** All four variables equal `gpt-5.5`, `gpt-5.5`, `gpt-5.5:medium`, `gpt-5.5:high` respectively, with zsh parse success.
+
+- [ ] **Step 2: Scenario MEDIUM — explicit override passthrough (L1+L2+L3)**
+
+Describe:
+- **Worker task:** Invoke the shell with `CONSENSUS_MODEL=gpt-5.4:high` exported before sourcing, then print the resulting `CONSENSUS_MODEL`. Also invoke with an unknown model string such as `CONSENSUS_MODEL=foo-model:custom` and print the result. Execution steps must include both runs and their stdout.
+- **Verifier categories:** (1) Correctness — both overrides are preserved byte-for-byte. (2) Completeness — no remap or warning is emitted to stderr. (3) Integration — the downstream variables that reference `CONSENSUS_MODEL` (e.g., when building consensus prompts) carry the override. (4) Docs alignment — `rlp-desk.md` describes `--consensus-model` as passthrough with a documented default. (5) Reasoning — justify why passthrough is the intended contract.
+- **PASS criteria:** `gpt-5.4:high` and `foo-model:custom` appear unchanged; stderr is silent about remapping.
+
+- [ ] **Step 3: Scenario CRITICAL — upgrade chain + security check (L1+L2+L3 + security)**
+
+Describe:
+- **Worker task:** In a Node one-off, import the `MODEL_UPGRADES` table from `src/node/runner/campaign-main-loop.mjs` (or inline-copy if import surface is not exposed) and assert the chain `gpt-5.5:medium → gpt-5.5:high → gpt-5.5:xhigh → BLOCKED`. Then verify the spark chain is untouched. Execution steps must include the exact Node invocation and output.
+- **Verifier categories:** (1) Correctness — the four assertions hold. (2) Completeness — no extra keys were introduced; legacy `gpt-5.4:*` keys are gone; spark keys remain exact. (3) Error-path — hitting `BLOCKED` is the terminal state; no silent key lookup fallback. (4) Security — confirm no shell expansion, template injection, or model-string concatenation was added to the hot path by this change set (spot-check `campaign-main-loop.mjs` diff). (5) Reasoning — one paragraph.
+- **PASS criteria:** All four assertions pass; spark chain intact; security spot-check clean.
+
+- [ ] **Step 4: Gate decision**
+
+If any scenario FAIL:
+- Fix the underlying code (not the test).
+- Re-run the failing scenario first; then re-run all three for clean slate.
+- Do not proceed to Task 9.
+
+If all three PASS, record the verdicts (even informally in the plan checkbox list) and move on.
+
+---
+
+## Task 9: Local file sync verification (CLAUDE.md §Local File Sync)
+
+**Files:** none modified; this is a sync-and-verify gate executed after commit approval.
+
+Per `CLAUDE.md`, after every commit that changes any `src/` file, sync ALL distributable files to `~/.claude/` and verify with `diff -q`.
+
+This task is listed for completeness but is executed AFTER user approval of the commit, which is outside the plan's automated scope.
+
+- [ ] **Step 1: Confirm sync list is current**
+
+Re-read the Local File Sync section of `CLAUDE.md` and confirm the list of files covers: `src/commands/rlp-desk.md`, `src/governance.md`, `src/model-upgrade-table.md`, `src/scripts/init_ralph_desk.zsh`, `src/scripts/run_ralph_desk.zsh`, `src/scripts/lib_ralph_desk.zsh`, `README.md`, `install.sh`, `docs/architecture.md`, `docs/getting-started.md`, `docs/protocol-reference.md`, `docs/TODO-verification-next.md`, `docs/internal/*`, `docs/blueprints/*`.
+
+- [ ] **Step 2: Sync all runtime files**
+
+Run:
+
+```bash
+cp src/commands/rlp-desk.md ~/.claude/commands/rlp-desk.md
+cp src/governance.md ~/.claude/ralph-desk/governance.md
+cp src/model-upgrade-table.md ~/.claude/ralph-desk/model-upgrade-table.md
+cp src/scripts/init_ralph_desk.zsh ~/.claude/ralph-desk/init_ralph_desk.zsh
+cp src/scripts/run_ralph_desk.zsh ~/.claude/ralph-desk/run_ralph_desk.zsh
+cp src/scripts/lib_ralph_desk.zsh ~/.claude/ralph-desk/lib_ralph_desk.zsh
+```
+
+- [ ] **Step 3: Sync reference docs**
+
+Run:
+
+```bash
+cp README.md ~/.claude/ralph-desk/README.md
+cp install.sh ~/.claude/ralph-desk/install.sh
+cp docs/architecture.md ~/.claude/ralph-desk/docs/architecture.md
+cp docs/getting-started.md ~/.claude/ralph-desk/docs/getting-started.md
+cp docs/protocol-reference.md ~/.claude/ralph-desk/docs/protocol-reference.md
+cp docs/TODO-verification-next.md ~/.claude/ralph-desk/docs/TODO-verification-next.md
+cp -r docs/internal/. ~/.claude/ralph-desk/docs/internal/
+cp -r docs/blueprints/. ~/.claude/ralph-desk/docs/blueprints/
+```
+
+- [ ] **Step 4: Verify each runtime file is byte-identical**
+
+Run:
+
+```bash
+diff -q src/commands/rlp-desk.md ~/.claude/commands/rlp-desk.md
+diff -q src/governance.md ~/.claude/ralph-desk/governance.md
+diff -q src/model-upgrade-table.md ~/.claude/ralph-desk/model-upgrade-table.md
+diff -q src/scripts/init_ralph_desk.zsh ~/.claude/ralph-desk/init_ralph_desk.zsh
+diff -q src/scripts/run_ralph_desk.zsh ~/.claude/ralph-desk/run_ralph_desk.zsh
+diff -q src/scripts/lib_ralph_desk.zsh ~/.claude/ralph-desk/lib_ralph_desk.zsh
+diff -q README.md ~/.claude/ralph-desk/README.md
+```
+
+Expected: no output from any of the seven `diff -q` commands. Any line of output means sync failed and must be repeated.
+
+---
+
+## Release workflow (outside plan automation, user approval gated)
+
+After Task 9 succeeds, the following release steps happen with explicit user approval at each stage, per `CLAUDE.md`:
+
+1. `ralplan` + `codex review` both reach zero issues.
+2. `git commit` with user-approved message.
+3. `git push -u origin feature/gpt-5-5-default`.
+4. Open PR, merge to `main` after approvals.
+5. `npm version` is already set to `0.10.0` inside `package.json` by Task 6; skip the implicit tag bump and commit the bump directly.
+6. `gh release create v0.10.0` with release notes limited to user-facing changes (see spec §Release notes draft).
+7. `npm publish`.
+8. Re-run Task 9 sync verification against the released tarball layout.
+
+Each step above is documented here as a reminder, not as an automated step. The engineer executing this plan must stop at Task 9 and request user approval before continuing.

package/docs/superpowers/specs/2026-04-24-gpt-5-5-default.md

@@ -0,0 +1,107 @@
+# Spec: Default Codex Model Migration gpt-5.4 → gpt-5.5
+
+**Date:** 2026-04-24
+**Version target:** 0.10.0 (minor bump)
+**Branch:** `feature/gpt-5-5-default`
+
+## Background
+
+Codex CLI 0.124.0 now ships with `gpt-5.5` as the default model in `~/.codex/config.toml` (replacing `gpt-5.4`). The user's reasoning effort default is `high`. rlp-desk's documented defaults, example commands, and upgrade tables still reference `gpt-5.4`, which creates drift between what users see in rlp-desk and what their Codex CLI actually runs.
+
+This spec migrates every rlp-desk default, example, and upgrade chain entry from `gpt-5.4` to `gpt-5.5`. User-supplied model flags are untouched — Codex CLI accepts any model name passthrough, so users who explicitly request `gpt-5.4:high` (or any other version) continue to work without rlp-desk interference.
+
+## Decisions
+
+| # | Decision | Rationale |
+|---|---|---|
+| D1 | Full replacement (`gpt-5.4` → `gpt-5.5` everywhere) | Codex CLI default changed; rlp-desk must mirror current reality. |
+| D2 | No auto-remap, no deprecation warning | User-supplied model names are passed through to Codex CLI as-is. rlp-desk is not a policy layer on model selection. |
+| D3 | Reasoning effort tiers (`low`/`medium`/`high`/`xhigh`) unchanged | Progressive upgrade philosophy preserved. Model name only. |
+| D4 | Minor bump 0.9.3 → 0.10.0 | Default behavior changes; semver-compliant. |
+
+## Scope
+
+### Files with `gpt-5.4` references (in-scope)
+
+| File | Change | Lines |
+|---|---|---|
+| `src/scripts/run_ralph_desk.zsh` | Default env values | 33, 35, 104, 106, 114-115, 2762, 2766 |
+| `src/node/run.mjs` | `consensusModel` / `finalConsensusModel` defaults | 14-15 |
+| `src/node/runner/campaign-main-loop.mjs` | Upgrade chain keys | 29-31 |
+| `src/model-upgrade-table.md` | Section header + full table | 21-28 |
+| `src/commands/rlp-desk.md` | Recommendation table, examples, help text | ~14 occurrences |
+| `package.json` | `version` field | — |
+
+### Out of scope (no changes)
+
+- `src/node/cli/command-builder.mjs` — no remap logic added.
+- `src/scripts/lib_ralph_desk.zsh` `parse_model_flag()` — accepts any model name as before.
+- `gpt-5.3-codex-spark` (GPT Pro) section in `model-upgrade-table.md` — distinct token limit, tracked separately.
+- Reasoning effort tier labels and progressive upgrade chain logic.
+- CLI flag names (`--worker-model`, `--consensus-model`, `--final-consensus-model`).
+
+## Release workflow
+
+Per `CLAUDE.md` gate rules (absolute):
+
+1. All source edits on `feature/gpt-5-5-default` branch.
+2. Self-verification gate triggered by changes to `src/commands/rlp-desk.md` and `src/scripts/run_ralph_desk.zsh`:
+   - LOW scenario: L1+L3
+   - MEDIUM scenario: L1+L2+L3
+   - CRITICAL scenario: L1+L2+L3 + security check
+   - All three must PASS before commit.
+3. `ralplan` + `codex review` must both reach 0 issues before merge.
+4. User approval for each: commit, version bump, merge, release, publish, local file sync.
+5. Local file sync of all 5 runtime files + reference docs per CLAUDE.md.
+6. Release notes: user-facing only (per `feedback_release_notes_scope.md`).
+
+## Release notes draft
+
+```
+### Changed
+- Default Codex model upgraded from `gpt-5.4` to `gpt-5.5` to match Codex CLI 0.124.0 defaults.
+- `WORKER_CODEX_MODEL`, `VERIFIER_CODEX_MODEL`, `CONSENSUS_MODEL`, `FINAL_CONSENSUS_MODEL` env defaults updated.
+- Upgrade chain in `model-upgrade-table.md` now references `gpt-5.5`.
+- Example commands in `/rlp-desk` recommendation table updated.
+
+### Unchanged
+- Users may still pass `--worker-model gpt-5.4:high` or any other model name explicitly; rlp-desk passes model flags through to Codex CLI without modification.
+- Reasoning effort tiers and progressive upgrade logic unchanged.
+- GPT Pro (`gpt-5.3-codex-spark`) path unchanged.
+```
+
+## Test plan
+
+### Self-verification scenarios (per CLAUDE.md §Self-Verification Gate)
+
+1. **LOW — default env value propagation.** Run `/rlp-desk run <slug>` with no model flags; confirm `WORKER_CODEX_MODEL=gpt-5.5` and `CONSENSUS_MODEL=gpt-5.5:medium` appear in the spawned command line.
+2. **MEDIUM — explicit override.** Run `/rlp-desk run <slug> --worker-model gpt-5.4:high`; confirm `gpt-5.4:high` is passed through unchanged to Codex CLI.
+3. **CRITICAL — upgrade chain.** Simulate a Worker failure; confirm the upgrade chain `gpt-5.5:medium → gpt-5.5:high → gpt-5.5:xhigh → BLOCKED` is followed.
+
+Each scenario: Worker (with `execution_steps`) → Verifier (5 categories, with `reasoning`) → PASS. Any FAIL blocks commit.
+
+### Regression sweep
+
+```bash
+grep -rn "gpt-5\.4" src/
+# Expected: zero hits.
+
+grep -rn "gpt-5\.4" docs/ --exclude-dir=superpowers/specs
+# Expected: zero hits. (This spec lives under superpowers/specs/ and intentionally mentions gpt-5.4.)
+```
+
+## Known follow-up (out of scope for v0.10.0)
+
+`tests/test_us004_self_verification.sh` has 4 pre-existing failures (PASS=42, FAIL=4) that surfaced during the post-migration full-suite sweep:
+- `AC4-boundary` / `AC13-happy`: `init_ralph_desk.zsh` lacks "no PRD found — treating as first-run" fallback language.
+- `AC13-boundary`: no note printed when `--mode` is provided but no PRD exists.
+- `AC13-runtime-happy`: `--mode improve` against a directory without a PRD does not exit 0 + print note + create a fresh PRD from template.
+
+Verified pre-existing via `git stash` baseline against pre-migration HEAD. These failures represent an unimplemented feature in `init_ralph_desk.zsh` (PRD template scaffolding for the no-PRD path), not a regression introduced by this migration. They will be tracked in a separate spec/PR (`init-no-prd-fallback`) targeting v0.10.1 or v0.11.0.
+
+## Rollback
+
+If a regression surfaces post-release:
+- Revert the `feature/gpt-5-5-default` merge commit.
+- Publish `0.10.1` with restored `gpt-5.4` defaults.
+- Users can always override via `--worker-model gpt-5.5:high` while rollback is in flight.

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "@ai-dev-methodologies/rlp-desk",
-  "version": "0.9.3",
+  "version": "0.10.0",
   "description": "Fresh-context iterative loops for Claude Code — autonomous task completion with independent verification",
   "scripts": {
     "postinstall": "node scripts/postinstall.js",

package/src/commands/rlp-desk.md

@@ -75,13 +75,13 @@ Ask about these items one by one (or in small groups):
 
    | Complexity | Worker | per-US Verifier | Final Verifier | Consensus |
    |------------|--------|-----------------|----------------|-----------|
-   | LOW | gpt-5.4:medium | sonnet | opus | final-only |
-   | MEDIUM | gpt-5.4:medium | opus | opus | final-only |
-   | HIGH | gpt-5.4:high | opus | opus | all |
-   | CRITICAL | gpt-5.4:high | opus | opus + human | all |
+   | LOW | gpt-5.5:medium | sonnet | opus | final-only |
+   | MEDIUM | gpt-5.5:medium | opus | opus | final-only |
+   | HIGH | gpt-5.5:high | opus | opus | all |
+   | CRITICAL | gpt-5.5:high | opus | opus + human | all |
 
    **Worker model selection** (cross-engine):
-   - **gpt-5.4:medium** — default recommendation (full context window, progressive upgrade handles harder US)
+   - **gpt-5.5:medium** — default recommendation (full context window, progressive upgrade handles harder US)
    - **spark:high** — only when US is small enough for spark's 100k context (single-file, AC count <= 4, simple logic). Do NOT use as primary recommendation — spark context window is too small for most tasks
 
    Present complexity score with evidence to the user, e.g.: "I rate this MEDIUM because: US count=4 (MEDIUM), file scope=2 (MEDIUM), logic=conditionals (MEDIUM), deps=none (LOW), impact=modify (MEDIUM). Highest=MEDIUM."
@@ -91,8 +91,8 @@ Ask about these items one by one (or in small groups):
    **If codex is NOT installed** — say: "Codex is not installed. Defaulting to claude-only Worker. Note: without a second engine, your Verifier shares the same perspective as the Worker — there is a risk of blind spots where both Worker and Verifier miss the same issue. To unlock cross-engine coverage: `npm install -g @openai/codex`"
 
 8. **Batch Capacity Check** — when verify-mode is batch and PRD is large:
-   - batch + spark + AC > 4 → warn "spark 100k context limit — switch to gpt-5.4 or split smaller"
-   - batch + gpt-5.4 + AC > 15 → warn "too many ACs for single batch — consider wave split (3-4 US per wave)"
+   - batch + spark + AC > 4 → warn "spark 100k context limit — switch to gpt-5.5 or split smaller"
+   - batch + gpt-5.5 + AC > 15 → warn "too many ACs for single batch — consider wave split (3-4 US per wave)"
    - per-us → no warning (US-level processing, no limit concern)
 9. **Verify Mode** — per-us (default) or batch. Ask: "Verify after each user story (per-us, recommended) or only after all stories are done (batch)?" Default recommendation: per-us for 2+ stories.
 10. **Consensus** — Ask: "Use cross-engine consensus? off (single engine), final-only (cross-engine on final verify only), or all (cross-engine on every verify). Requires codex CLI." Default: off. Recommended: final-only when codex is installed.
@@ -164,26 +164,26 @@ Tell the user:
    Available run commands (copy the one you want):
 
    # ★ Recommended: cross-engine + final-consensus (full context + blind-spot coverage):
-   /rlp-desk run <actual-slug> --mode tmux --worker-model gpt-5.4:medium --consensus final-only --debug
+   /rlp-desk run <actual-slug> --mode tmux --worker-model gpt-5.5:medium --consensus final-only --debug
 
    # Small tasks only (single-file, AC <= 4, simple logic — spark 100k context limit):
    /rlp-desk run <actual-slug> --mode tmux --worker-model spark:high --consensus final-only --debug
 
    # Critical (full consensus on every verify):
-   /rlp-desk run <actual-slug> --mode tmux --worker-model gpt-5.4:high --consensus all --debug
+   /rlp-desk run <actual-slug> --mode tmux --worker-model gpt-5.5:high --consensus all --debug
 
    # Claude-only:
    /rlp-desk run <actual-slug> --debug
 
    # Full options reference:
    #   --mode agent|tmux                      (default: agent)
-   #   --worker-model MODEL                   haiku|sonnet|opus or gpt-5.4:high|spark:high (default: haiku)
+   #   --worker-model MODEL                   haiku|sonnet|opus or gpt-5.5:high|spark:high (default: haiku)
    #   --lock-worker-model                    disable auto model upgrade
    #   --verifier-model MODEL                 per-US verifier (default: sonnet)
    #   --final-verifier-model MODEL           final ALL verifier (default: opus)
    #   --consensus off|all|final-only         cross-engine consensus (default: off)
-   #   --consensus-model MODEL                per-US cross-verifier (default: gpt-5.4:medium)
-   #   --final-consensus-model MODEL          final cross-verifier (default: gpt-5.4:high)
+   #   --consensus-model MODEL                per-US cross-verifier (default: gpt-5.5:medium)
+   #   --final-consensus-model MODEL          final cross-verifier (default: gpt-5.5:high)
    #   --verify-mode per-us|batch             (default: per-us)
    #   --cb-threshold N                       (default: 6)
    #   --max-iter N                           (default: 100)
@@ -240,16 +240,16 @@ Tell the user:
 
 Options (parse from `$ARGUMENTS`):
 - `--mode agent|tmux` (default: `agent`) — execution mode
-- `--worker-model MODEL` (default: `haiku`) — Worker model. Format: `model` = claude engine, `model:reasoning` = codex engine. Examples: `haiku`, `sonnet`, `opus`, `spark:high`, `gpt-5.4:high`. Parsed by `parse_model_flag()` which auto-splits engine/model/reasoning.
-- `--lock-worker-model` — disable automatic model upgrade on failure (check_model_upgrade). Worker stays on the specified model regardless of consecutive failures.
+- `--worker-model MODEL` (default: `haiku`) — Worker model. Format: `model` = claude engine, `model:reasoning` = codex engine. Examples: `haiku`, `sonnet`, `opus`, `spark:high`, `gpt-5.5:high`. Parsed by `parse_model_flag()` which auto-splits engine/model/reasoning.
+- `--lock-worker-model` — disable automatic model upgrade on failure. Worker stays on the specified model regardless of consecutive failures.
 - `--verifier-model MODEL` (default: `sonnet`) — per-US verification model. Campaign-fixed (no progressive upgrade). Lighter than final verifier.
 - `--final-verifier-model MODEL` (default: `opus`) — final ALL verification model. Independent from per-US verifier. Used only for the final full-AC verify pass.
 - `--consensus off|all|final-only` (default: `off`) — cross-engine consensus verification mode.
   - `off`: single-engine verification only
   - `all`: cross-engine consensus on every verify (per-US and final)
   - `final-only`: cross-engine consensus only on the final ALL verify
-- `--consensus-model MODEL` (default: `gpt-5.4:medium`) — per-US cross-verifier model. Lighter weight for cost efficiency.
-- `--final-consensus-model MODEL` (default: `gpt-5.4:high`) — final cross-verifier model. Stricter. Note: spark is not allowed here (100k output limit).
+- `--consensus-model MODEL` (default: `gpt-5.5:medium`) — per-US cross-verifier model. Lighter weight for cost efficiency.
+- `--final-consensus-model MODEL` (default: `gpt-5.5:high`) — final cross-verifier model. Stricter. Note: spark is not allowed here (100k output limit).
 - `--verify-mode per-us|batch` (default: `per-us`) — verification strategy
   - `per-us`: verify after each US, then final full verify of all AC
   - `batch`: verify only after all US done (legacy behavior)
@@ -292,8 +292,8 @@ VERIFIER_MODEL=<--verifier-model value, default: sonnet> \
 FINAL_VERIFIER_MODEL=<--final-verifier-model value, default: opus> \
 VERIFY_MODE=<--verify-mode value, default: per-us> \
 CONSENSUS_MODE=<--consensus value, default: off> \
-CONSENSUS_MODEL=<--consensus-model value, default: gpt-5.4:medium> \
-FINAL_CONSENSUS_MODEL=<--final-consensus-model value, default: gpt-5.4:high> \
+CONSENSUS_MODEL=<--consensus-model value, default: gpt-5.5:medium> \
+FINAL_CONSENSUS_MODEL=<--final-consensus-model value, default: gpt-5.5:high> \
 CB_THRESHOLD=<--cb-threshold value, default: 6> \
 ITER_TIMEOUT=<--iter-timeout value, default: 600> \
 DEBUG=<1 if --debug, else 0> \
@@ -473,8 +473,8 @@ Bash("codex exec --model <codex_model> --reasoning-effort <codex_reasoning> <ful
 **⑦b Consensus Verification** (when `--consensus` is `all`, or `final-only` and scope is ALL):
 After the primary verifier runs, run a cross-engine second verifier:
 - Determine cross-verifier model based on scope:
-  - per-US verify → use `--consensus-model` (default: gpt-5.4:medium)
-  - final ALL verify → use `--final-consensus-model` (default: gpt-5.4:high)
+  - per-US verify → use `--consensus-model` (default: gpt-5.5:medium)
+  - final ALL verify → use `--final-consensus-model` (default: gpt-5.5:high)
 - If primary engine is claude → cross-verifier uses codex (the consensus model)
 - If primary engine is codex → cross-verifier uses claude `opus` (fixed)
 - Both produce `verify-verdict.json` (Leader renames to `verify-verdict-claude.json` and `verify-verdict-codex.json`)
@@ -732,13 +732,13 @@ Example:
 
 Run options:
   --mode agent|tmux                    Execution mode (default: agent)
-  --worker-model MODEL                 Worker model: haiku|sonnet|opus or gpt-5.4:high|spark:high (default: haiku)
+  --worker-model MODEL                 Worker model: haiku|sonnet|opus or gpt-5.5:high|spark:high (default: haiku)
   --lock-worker-model                  Disable auto model upgrade on failure
   --verifier-model MODEL               per-US verifier (default: sonnet)
   --final-verifier-model MODEL         Final ALL verifier (default: opus)
   --consensus off|all|final-only       Cross-engine consensus (default: off)
-  --consensus-model MODEL              per-US cross-verifier (default: gpt-5.4:medium)
-  --final-consensus-model MODEL        Final cross-verifier (default: gpt-5.4:high)
+  --consensus-model MODEL              per-US cross-verifier (default: gpt-5.5:medium)
+  --final-consensus-model MODEL        Final cross-verifier (default: gpt-5.5:high)
   --verify-mode per-us|batch           Verification strategy (default: per-us)
   --cb-threshold N                     Consecutive failures before BLOCKED (default: 6)
   --max-iter N                         Max iterations (default: 100)

package/src/governance.md

@@ -14,7 +14,7 @@ The Leader orchestrates, while Worker/Verifier run in isolated fresh contexts ev
 - **Worker must NEVER modify Claude Code settings** (settings.json, settings.local.json). Permission prompts must be reported as blocked, not bypassed by editing settings.
 - **Verifier is independent**: The Verifier judges based on evidence alone, without knowledge of the Worker's reasoning process.
 - **Sentinels are Leader-owned**: Only the Leader writes COMPLETE/BLOCKED sentinels.
-- **Supported engines**: claude (default; models: haiku, sonnet, opus) and codex (opt-in via `--worker-model spark:high` or `--worker-model gpt-5.4:high`).
+- **Supported engines**: claude (default; models: haiku, sonnet, opus) and codex (opt-in via `--worker-model spark:high` or `--worker-model gpt-5.5:high`).
 
 ## 1a. Iron Laws
 
@@ -300,11 +300,11 @@ The Leader decides each iteration. Decision criteria:
 
 ### Codex (opt-in engine)
 
-Model routing uses `--worker-model` and `--verifier-model` with codex format: `spark:high` or `gpt-5.4:high`.
+Model routing uses `--worker-model` and `--verifier-model` with codex format: `spark:high` or `gpt-5.5:high`.
 
 ```
 --worker-model spark:high        # codex worker, spark model, high reasoning
---verifier-model gpt-5.4:high    # codex verifier, gpt-5.4, high reasoning
+--verifier-model gpt-5.5:high    # codex verifier, gpt-5.5, high reasoning
 ```
 
 `parse_model_flag()` auto-detects engine from the model name: plain names (haiku, sonnet, opus) = claude; `name:reasoning` format = codex. Claude is the default engine; codex is explicitly opt-in.
@@ -331,7 +331,7 @@ Agent(
 )
 ```
 
-If `--worker-model` or `--verifier-model` uses codex format (e.g., `spark:high`, `gpt-5.4:high`) (opt-in):
+If `--worker-model` or `--verifier-model` uses codex format (e.g., `spark:high`, `gpt-5.5:high`) (opt-in):
 ```
 # Worker or Verifier (codex engine)
 Bash("codex -m <codex_model> -c model_reasoning_effort=<codex_reasoning> --dangerously-bypass-approvals-and-sandbox <prompt>")
@@ -377,7 +377,7 @@ claude -p "$(cat /path/to/prompt.md)" \
 When `WORKER_ENGINE=codex` or `VERIFIER_ENGINE=codex`, the `codex` CLI is used instead:
 ```bash
 # codex engine (opt-in)
-codex -m gpt-5.4 \
+codex -m gpt-5.5 \
   -c model_reasoning_effort="high" \
   --dangerously-bypass-approvals-and-sandbox \
   "$(cat /path/to/prompt.md)"
@@ -569,9 +569,9 @@ Worker completes US → signal verify
 
 | Scenario | Primary verifier | Cross verifier |
 |----------|-----------------|----------------|
-| per-US, primary=claude | `--verifier-model` (sonnet) | `--consensus-model` (gpt-5.4:medium) |
+| per-US, primary=claude | `--verifier-model` (sonnet) | `--consensus-model` (gpt-5.5:medium) |
 | per-US, primary=codex | `--verifier-model` | claude opus (fixed) |
-| final, primary=claude | `--final-verifier-model` (opus) | `--final-consensus-model` (gpt-5.4:high) |
+| final, primary=claude | `--final-verifier-model` (opus) | `--final-consensus-model` (gpt-5.5:high) |
 | final, primary=codex | `--final-verifier-model` | claude opus (fixed) |
 
 - Both must pass. No engine priority.

package/src/model-upgrade-table.md

@@ -18,14 +18,14 @@ CB default: 6. Override: `--cb-threshold N`. Worker only — Verifier fixed at c
 | HIGH | gpt-5.3-codex-spark:high | gpt-5.3-codex-spark:xhigh | gpt-5.3-codex-spark:xhigh | BLOCKED |
 | CRITICAL | gpt-5.3-codex-spark:xhigh | gpt-5.3-codex-spark:xhigh | gpt-5.3-codex-spark:xhigh | BLOCKED |
 
-## Non-Pro (gpt-5.4)
+## Non-Pro (gpt-5.5)
 
 | Complexity | 1-2 | 3-4 | 5-6 | 7+ |
 |------------|-----|-----|-----|-----|
-| LOW | gpt-5.4:low | gpt-5.4:medium | gpt-5.4:high | BLOCKED |
-| MEDIUM | gpt-5.4:medium | gpt-5.4:high | gpt-5.4:xhigh | BLOCKED |
-| HIGH | gpt-5.4:high | gpt-5.4:xhigh | gpt-5.4:xhigh | BLOCKED |
-| CRITICAL | gpt-5.4:xhigh | gpt-5.4:xhigh | gpt-5.4:xhigh | BLOCKED |
+| LOW | gpt-5.5:low | gpt-5.5:medium | gpt-5.5:high | BLOCKED |
+| MEDIUM | gpt-5.5:medium | gpt-5.5:high | gpt-5.5:xhigh | BLOCKED |
+| HIGH | gpt-5.5:high | gpt-5.5:xhigh | gpt-5.5:xhigh | BLOCKED |
+| CRITICAL | gpt-5.5:xhigh | gpt-5.5:xhigh | gpt-5.5:xhigh | BLOCKED |
 
 ## Claude-only
 

package/src/node/run.mjs

@@ -11,8 +11,8 @@ const RUN_DEFAULTS = {
   verifierModel: 'sonnet',
   finalVerifierModel: 'opus',
   consensusMode: 'off',
-  consensusModel: 'gpt-5.4:medium',
-  finalConsensusModel: 'gpt-5.4:high',
+  consensusModel: 'gpt-5.5:medium',
+  finalConsensusModel: 'gpt-5.5:high',
   verifyMode: 'per-us',
   cbThreshold: 6,
   maxIterations: 100,

package/src/node/runner/campaign-main-loop.mjs

@@ -26,9 +26,9 @@ const execFileAsync = promisify(execFile);
 const REQUIRED_SCAFFOLD_NAMES = ['workerPrompt', 'verifierPrompt', 'memoryFile', 'prdFile', 'testSpecFile'];
 const CLAUDE_MODELS = new Set(['haiku', 'sonnet', 'opus']);
 const MODEL_UPGRADES = {
-  'gpt-5.4:medium': 'gpt-5.4:high',
-  'gpt-5.4:high': 'gpt-5.4:xhigh',
-  'gpt-5.4:xhigh': 'BLOCKED',
+  'gpt-5.5:medium': 'gpt-5.5:high',
+  'gpt-5.5:high': 'gpt-5.5:xhigh',
+  'gpt-5.5:xhigh': 'BLOCKED',
   'gpt-5.3-codex-spark:medium': 'gpt-5.3-codex-spark:high',
   'gpt-5.3-codex-spark:high': 'gpt-5.3-codex-spark:xhigh',
   'gpt-5.3-codex-spark:xhigh': 'BLOCKED',