axstack 0.20.8 → 0.20.10

package/docs/workflows.md

@@ -51,7 +51,7 @@ role.
 | Preset | Author | Ordered peer reviewers | Astra / Fable advisers | Auditor |
 | --- | --- | --- | --- | --- |
 | `mixed` | Sol medium | Sol medium; Opus medium | Astra high / Fable high | Luna max |
-| `codex-only` | Sol medium | Sol medium; Terra xhigh | Astra high / unavailable | Luna max |
+| `codex-only` | Sol medium | Sol medium; Luna xhigh | Astra high / unavailable | Luna max |
 | `claude-only` | Opus medium | Opus medium; Sonnet xhigh | unavailable / Fable high | Sonnet xhigh |
 
 The installed `<skills-dir>/axstack/roles.json` adds the selected preset name:

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "axstack",
-  "version": "0.20.8",
+  "version": "0.20.10",
   "description": "Axstack installer and setup CLI: installs owned chat skills and role data, configures supported harness settings, and checks Orca capabilities.",
   "keywords": [
     "claude-code",

package/profiles/presets/claude-only.json

@@ -23,7 +23,7 @@
       "id": "axstack-owner",
       "name": "Axstack PR owner",
       "provider": "claude",
-      "model": "claude-opus-5",
+      "model": "claude-opus-5-5",
       "modeId": "bypassPermissions",
       "thinkingOptionId": "high",
       "notes": "Persistent PR owner: one owner per PR, accountable for candidate, fixes, verification evidence, and monitoring. May delegate coding but never edits a worker-owned candidate concurrently. Launches eligible independent reviewers."
@@ -32,7 +32,7 @@
       "id": "axstack-author",
       "name": "Axstack author",
       "provider": "claude",
-      "model": "claude-opus-5",
+      "model": "claude-opus-5-5",
       "modeId": "bypassPermissions",
       "thinkingOptionId": "medium",
       "notes": "Ordinary implementation and repairs. Uses strict red-green-refactor and remains the exclusive writer for a candidate. Validate configured availability at launch; hold affected work without fallback."
@@ -41,7 +41,7 @@
       "id": "axstack-reviewer-primary",
       "name": "Axstack reviewer (primary)",
       "provider": "claude",
-      "model": "claude-opus-5",
+      "model": "claude-opus-5-5",
       "modeId": "bypassPermissions",
       "thinkingOptionId": "medium",
       "notes": "Primary reviewer in the ordered claude-only peer pair: Opus medium followed by Sonnet xhigh. Never author or owner; review exact SHA and base across all six angles and acceptance. Peer first pass stays isolated."
@@ -68,7 +68,7 @@
       "id": "axstack-research-requirements",
       "name": "Axstack research requirements",
       "provider": "claude",
-      "model": "claude-opus-5",
+      "model": "claude-opus-5-5",
       "modeId": "bypassPermissions",
       "thinkingOptionId": "medium",
       "notes": "Research requirements analyst: scopes bounded questions and acceptance for a research task. Validate configured availability at launch; hold affected work without fallback."
@@ -77,7 +77,7 @@
       "id": "axstack-research-code",
       "name": "Axstack research code",
       "provider": "claude",
-      "model": "claude-opus-5",
+      "model": "claude-opus-5-5",
       "modeId": "bypassPermissions",
       "thinkingOptionId": "medium",
       "notes": "Research code investigator: verifies behavior against inspected code and executable evidence. Validate configured availability at launch; hold affected work without fallback."
@@ -167,10 +167,10 @@
       "id": "axstack-debug-investigator-1",
       "name": "Axstack debug investigator 1",
       "provider": "claude",
-      "model": "claude-opus-5",
+      "model": "claude-opus-5-5",
       "modeId": "bypassPermissions",
       "thinkingOptionId": "medium",
-      "notes": "Debug investigator seat 1. Dispatched only by axstack-debug at L1 with the shared evidence packet and one distinct brief; never reads another investigator's output. Works in its own disposable worktree at the pinned revision plus the recorded dirty patch; may instrument there for probes; never commits, pushes, publishes, or creates children. Returns one receipt per brief. Independence comes from brief isolation, not model diversity. This preset repeats claude-opus-5 at medium effort because it has fewer model families."
+      "notes": "Debug investigator seat 1. Dispatched only by axstack-debug at L1 with the shared evidence packet and one distinct brief; never reads another investigator's output. Works in its own disposable worktree at the pinned revision plus the recorded dirty patch; may instrument there for probes; never commits, pushes, publishes, or creates children. Returns one receipt per brief. Independence comes from brief isolation, not model diversity. This preset repeats claude-opus-5-5 at medium effort because it has fewer model families."
     },
     {
       "id": "axstack-debug-investigator-2",
@@ -185,10 +185,10 @@
       "id": "axstack-debug-investigator-3",
       "name": "Axstack debug investigator 3",
       "provider": "claude",
-      "model": "claude-opus-5",
+      "model": "claude-opus-5-5",
       "modeId": "bypassPermissions",
       "thinkingOptionId": "high",
-      "notes": "Debug investigator seat 3. Dispatched only by axstack-debug at L1 with the shared evidence packet and one distinct brief; never reads another investigator's output. Works in its own disposable worktree at the pinned revision plus the recorded dirty patch; may instrument there for probes; never commits, pushes, publishes, or creates children. Returns one receipt per brief. Independence comes from brief isolation, not model diversity. This preset repeats claude-opus-5 at high effort because it has fewer model families."
+      "notes": "Debug investigator seat 3. Dispatched only by axstack-debug at L1 with the shared evidence packet and one distinct brief; never reads another investigator's output. Works in its own disposable worktree at the pinned revision plus the recorded dirty patch; may instrument there for probes; never commits, pushes, publishes, or creates children. Returns one receipt per brief. Independence comes from brief isolation, not model diversity. This preset repeats claude-opus-5-5 at high effort because it has fewer model families."
     },
     {
       "id": "axstack-debug-investigator-4",

package/profiles/presets/codex-only.json

@@ -23,7 +23,7 @@
       "id": "axstack-owner",
       "name": "Axstack PR owner",
       "provider": "codex",
-      "model": "gpt-5.6-sol",
+      "model": "gpt-6-sol",
       "modeId": "full-access",
       "thinkingOptionId": "high",
       "notes": "Persistent PR owner: one owner per PR, accountable for candidate, fixes, verification evidence, and monitoring. May delegate coding but never edits a worker-owned candidate concurrently. Launches eligible independent reviewers."
@@ -32,7 +32,7 @@
       "id": "axstack-author",
       "name": "Axstack author",
       "provider": "codex",
-      "model": "gpt-5.6-sol",
+      "model": "gpt-6-sol",
       "modeId": "full-access",
       "thinkingOptionId": "medium",
       "notes": "Ordinary implementation and repairs. Uses strict red-green-refactor and remains the exclusive writer for a candidate. Validate configured availability at launch; hold affected work without fallback."
@@ -41,25 +41,25 @@
       "id": "axstack-reviewer-primary",
       "name": "Axstack reviewer (primary)",
       "provider": "codex",
-      "model": "gpt-5.6-sol",
+      "model": "gpt-6-sol",
       "modeId": "full-access",
       "thinkingOptionId": "medium",
-      "notes": "Primary reviewer in the ordered codex-only peer pair: Sol medium followed by Terra xhigh. Never author or owner; review exact SHA and base across all six angles and acceptance. Peer first pass stays isolated."
+      "notes": "Primary reviewer in the ordered codex-only peer pair: Sol medium followed by Luna xhigh. Never author or owner; review exact SHA and base across all six angles and acceptance. Peer first pass stays isolated."
     },
     {
       "id": "axstack-reviewer-secondary",
       "name": "Axstack reviewer (secondary)",
       "provider": "codex",
-      "model": "gpt-5.6-terra",
+      "model": "gpt-6-luna",
       "modeId": "full-access",
       "thinkingOptionId": "xhigh",
-      "notes": "Secondary reviewer in the ordered codex-only peer pair: Sol medium followed by Terra xhigh. Eligible authored reviewer for a Sol-authored candidate. Never author or owner; review exact SHA and base across all six angles and acceptance. Peer first pass stays isolated."
+      "notes": "Secondary reviewer in the ordered codex-only peer pair: Sol medium followed by Luna xhigh. Eligible authored reviewer for a Sol-authored candidate. Never author or owner; review exact SHA and base across all six angles and acceptance. Peer first pass stays isolated."
     },
     {
       "id": "axstack-checker",
       "name": "Axstack tracker checker",
       "provider": "codex",
-      "model": "gpt-5.6-luna",
+      "model": "gpt-6-luna",
       "modeId": "full-access",
       "thinkingOptionId": "low",
       "notes": "Report-only discrepancy checker for the selected external tracker (Linear or GitHub Issues). Never mutates the tracker; the driver independently verifies evidence before applying updates. A null model means explicit user selection is required before dispatch and must never launch a provider default."
@@ -77,7 +77,7 @@
       "id": "axstack-research-code",
       "name": "Axstack research code",
       "provider": "codex",
-      "model": "gpt-5.6-sol",
+      "model": "gpt-6-sol",
       "modeId": "full-access",
       "thinkingOptionId": "medium",
       "notes": "Research code investigator: verifies behavior against inspected code and executable evidence. Validate configured availability at launch; hold affected work without fallback."
@@ -86,7 +86,7 @@
       "id": "axstack-research-web",
       "name": "Axstack research web",
       "provider": "codex",
-      "model": "gpt-5.6-terra",
+      "model": "gpt-6-sol",
       "modeId": "full-access",
       "thinkingOptionId": "low",
       "notes": "Research web reader: gathers primary-source facts efficiently. Validate configured availability at launch; hold affected work without fallback."
@@ -113,7 +113,7 @@
       "id": "axstack-explainer",
       "name": "Axstack explainer",
       "provider": "codex",
-      "model": "gpt-5.6-sol",
+      "model": "gpt-6-sol",
       "modeId": "full-access",
       "thinkingOptionId": "high",
       "notes": "Complex visual explanation author: traces systems, changes, and implementation gaps in requested artifacts and verifies rendered behavior where applicable. Validate configured availability at launch; hold affected work without fallback."
@@ -122,7 +122,7 @@
       "id": "axstack-explainer-review",
       "name": "Axstack explainer reviewer",
       "provider": "codex",
-      "model": "gpt-5.6-luna",
+      "model": "gpt-6-luna",
       "modeId": "full-access",
       "thinkingOptionId": "max",
       "notes": "Independent visual explanation reviewer: checks the exact artifact for source fidelity and rendered behavior where warranted. Any artifact change invalidates its review. Validate configured availability at launch; hold affected work without fallback."
@@ -131,7 +131,7 @@
       "id": "axstack-explore-codebase",
       "name": "Axstack codebase explorer",
       "provider": "codex",
-      "model": "gpt-5.6-terra",
+      "model": "gpt-6-sol",
       "modeId": "full-access",
       "thinkingOptionId": "xhigh",
       "notes": "Codebase mapper: explores repository structure and interfaces for research and handoff context. Validate configured availability at launch; hold affected work without fallback."
@@ -140,7 +140,7 @@
       "id": "axstack-explore-execution",
       "name": "Axstack execution explorer",
       "provider": "codex",
-      "model": "gpt-5.6-terra",
+      "model": "gpt-6-sol",
       "modeId": "full-access",
       "thinkingOptionId": "low",
       "notes": "Execution explorer: runs bounded checks of runtime behavior where authorized. Validate configured availability at launch; hold affected work without fallback."
@@ -149,7 +149,7 @@
       "id": "axstack-monitor",
       "name": "Axstack monitor",
       "provider": "codex",
-      "model": "gpt-5.6-terra",
+      "model": "gpt-6-sol",
       "modeId": "full-access",
       "thinkingOptionId": "low",
       "notes": "Optional independent read-only observer for a standalone PR watch. Reads GitHub, feedback, and checks, persists event IDs, and wakes the owner only for a new actionable event. Never sends, authors, reviews, replies, or acts as either reusable PR manager. Healthy snapshots stay quiet."
@@ -158,7 +158,7 @@
       "id": "axstack-auditor",
       "name": "Axstack auditor",
       "provider": "codex",
-      "model": "gpt-5.6-luna",
+      "model": "gpt-6-luna",
       "modeId": "full-access",
       "thinkingOptionId": "max",
       "notes": "Read-only end-of-run and checkpoint auditor. Collects scope and outcome evidence with counts and denominators and reports PASS, FAIL, or UNKNOWN without inventing numbers. Never edits, merges, activates, or audits itself."
@@ -167,37 +167,37 @@
       "id": "axstack-debug-investigator-1",
       "name": "Axstack debug investigator 1",
       "provider": "codex",
-      "model": "gpt-5.6-sol",
+      "model": "gpt-6-sol",
       "modeId": "full-access",
       "thinkingOptionId": "medium",
-      "notes": "Debug investigator seat 1. Dispatched only by axstack-debug at L1 with the shared evidence packet and one distinct brief; never reads another investigator's output. Works in its own disposable worktree at the pinned revision plus the recorded dirty patch; may instrument there for probes; never commits, pushes, publishes, or creates children. Returns one receipt per brief. Independence comes from brief isolation, not model diversity. This preset repeats gpt-5.6-sol at medium effort because it has fewer model families."
+      "notes": "Debug investigator seat 1. Dispatched only by axstack-debug at L1 with the shared evidence packet and one distinct brief; never reads another investigator's output. Works in its own disposable worktree at the pinned revision plus the recorded dirty patch; may instrument there for probes; never commits, pushes, publishes, or creates children. Returns one receipt per brief. Independence comes from brief isolation, not model diversity. This preset repeats gpt-6-sol at medium effort because it has fewer model families."
     },
     {
       "id": "axstack-debug-investigator-2",
       "name": "Axstack debug investigator 2",
       "provider": "codex",
-      "model": "gpt-5.6-terra",
+      "model": "gpt-6-sol",
       "modeId": "full-access",
       "thinkingOptionId": "low",
-      "notes": "Debug investigator seat 2. Dispatched only by axstack-debug at L1 with the shared evidence packet and one distinct brief; never reads another investigator's output. Works in its own disposable worktree at the pinned revision plus the recorded dirty patch; may instrument there for probes; never commits, pushes, publishes, or creates children. Returns one receipt per brief. Independence comes from brief isolation, not model diversity. This preset repeats gpt-5.6-terra at low effort because it has fewer model families."
+      "notes": "Debug investigator seat 2. Dispatched only by axstack-debug at L1 with the shared evidence packet and one distinct brief; never reads another investigator's output. Works in its own disposable worktree at the pinned revision plus the recorded dirty patch; may instrument there for probes; never commits, pushes, publishes, or creates children. Returns one receipt per brief. Independence comes from brief isolation, not model diversity. This preset repeats gpt-6-sol at low effort because it has fewer model families."
     },
     {
       "id": "axstack-debug-investigator-3",
       "name": "Axstack debug investigator 3",
       "provider": "codex",
-      "model": "gpt-5.6-sol",
+      "model": "gpt-6-sol",
       "modeId": "full-access",
       "thinkingOptionId": "high",
-      "notes": "Debug investigator seat 3. Dispatched only by axstack-debug at L1 with the shared evidence packet and one distinct brief; never reads another investigator's output. Works in its own disposable worktree at the pinned revision plus the recorded dirty patch; may instrument there for probes; never commits, pushes, publishes, or creates children. Returns one receipt per brief. Independence comes from brief isolation, not model diversity. This preset repeats gpt-5.6-sol at high effort because it has fewer model families."
+      "notes": "Debug investigator seat 3. Dispatched only by axstack-debug at L1 with the shared evidence packet and one distinct brief; never reads another investigator's output. Works in its own disposable worktree at the pinned revision plus the recorded dirty patch; may instrument there for probes; never commits, pushes, publishes, or creates children. Returns one receipt per brief. Independence comes from brief isolation, not model diversity. This preset repeats gpt-6-sol at high effort because it has fewer model families."
     },
     {
       "id": "axstack-debug-investigator-4",
       "name": "Axstack debug investigator 4",
       "provider": "codex",
-      "model": "gpt-5.6-terra",
+      "model": "gpt-6-sol",
       "modeId": "full-access",
       "thinkingOptionId": "xhigh",
-      "notes": "Debug investigator seat 4. Dispatched only by axstack-debug at L1 with the shared evidence packet and one distinct brief; never reads another investigator's output. Works in its own disposable worktree at the pinned revision plus the recorded dirty patch; may instrument there for probes; never commits, pushes, publishes, or creates children. Returns one receipt per brief. Independence comes from brief isolation, not model diversity. This preset repeats gpt-5.6-terra at xhigh effort because it has fewer model families."
+      "notes": "Debug investigator seat 4. Dispatched only by axstack-debug at L1 with the shared evidence packet and one distinct brief; never reads another investigator's output. Works in its own disposable worktree at the pinned revision plus the recorded dirty patch; may instrument there for probes; never commits, pushes, publishes, or creates children. Returns one receipt per brief. Independence comes from brief isolation, not model diversity. This preset repeats gpt-6-sol at xhigh effort because it has fewer model families."
     },
     {
       "id": "axstack-arena-judge-astra",

package/profiles/presets/mixed.json

@@ -23,7 +23,7 @@
       "id": "axstack-owner",
       "name": "Axstack PR owner",
       "provider": "claude",
-      "model": "claude-opus-5",
+      "model": "claude-opus-5-5",
       "modeId": "bypassPermissions",
       "thinkingOptionId": "medium",
       "notes": "Persistent PR owner: one owner per PR, accountable for candidate, fixes, verification evidence, and monitoring. May delegate coding but never edits a worker-owned candidate concurrently. Launches eligible independent reviewers."
@@ -32,7 +32,7 @@
       "id": "axstack-author",
       "name": "Axstack author",
       "provider": "codex",
-      "model": "gpt-5.6-sol",
+      "model": "gpt-6-sol",
       "modeId": "full-access",
       "thinkingOptionId": "medium",
       "notes": "Ordinary implementation and repairs. Uses strict red-green-refactor and remains the exclusive writer for a candidate. Validate configured availability at launch; hold affected work without fallback."
@@ -41,7 +41,7 @@
       "id": "axstack-reviewer-primary",
       "name": "Axstack reviewer (primary)",
       "provider": "codex",
-      "model": "gpt-5.6-sol",
+      "model": "gpt-6-sol",
       "modeId": "full-access",
       "thinkingOptionId": "medium",
       "notes": "Primary reviewer in the ordered mixed peer pair: Sol medium followed by Opus medium. Eligible authored reviewer for an Opus-authored candidate. Never author or owner; review exact SHA and base across all six angles and acceptance. Peer first pass stays isolated."
@@ -50,7 +50,7 @@
       "id": "axstack-reviewer-secondary",
       "name": "Axstack reviewer (secondary)",
       "provider": "claude",
-      "model": "claude-opus-5",
+      "model": "claude-opus-5-5",
       "modeId": "bypassPermissions",
       "thinkingOptionId": "medium",
       "notes": "Secondary reviewer in the ordered mixed peer pair: Sol medium followed by Opus medium. Eligible authored reviewer for a Sol-authored candidate. Never author or owner; review exact SHA and base across all six angles and acceptance. Peer first pass stays isolated."
@@ -68,7 +68,7 @@
       "id": "axstack-research-requirements",
       "name": "Axstack research requirements",
       "provider": "claude",
-      "model": "claude-opus-5",
+      "model": "claude-opus-5-5",
       "modeId": "bypassPermissions",
       "thinkingOptionId": "medium",
       "notes": "Research requirements analyst: scopes bounded questions and acceptance for a research task. Validate configured availability at launch; hold affected work without fallback."
@@ -77,7 +77,7 @@
       "id": "axstack-research-code",
       "name": "Axstack research code",
       "provider": "codex",
-      "model": "gpt-5.6-sol",
+      "model": "gpt-6-sol",
       "modeId": "full-access",
       "thinkingOptionId": "medium",
       "notes": "Research code investigator: verifies behavior against inspected code and executable evidence. Validate configured availability at launch; hold affected work without fallback."
@@ -86,7 +86,7 @@
       "id": "axstack-research-web",
       "name": "Axstack research web",
       "provider": "claude",
-      "model": "claude-opus-5",
+      "model": "claude-opus-5-5",
       "modeId": "bypassPermissions",
       "thinkingOptionId": "low",
       "notes": "Research web reader: gathers primary-source facts efficiently. Validate configured availability at launch; hold affected work without fallback."
@@ -122,7 +122,7 @@
       "id": "axstack-explainer-review",
       "name": "Axstack explainer reviewer",
       "provider": "codex",
-      "model": "gpt-5.6-luna",
+      "model": "gpt-6-luna",
       "modeId": "full-access",
       "thinkingOptionId": "max",
       "notes": "Independent visual explanation reviewer: checks the exact artifact for source fidelity and rendered behavior where warranted. Any artifact change invalidates its review. Validate configured availability at launch; hold affected work without fallback."
@@ -140,7 +140,7 @@
       "id": "axstack-explore-execution",
       "name": "Axstack execution explorer",
       "provider": "codex",
-      "model": "gpt-5.6-terra",
+      "model": "gpt-6-sol",
       "modeId": "full-access",
       "thinkingOptionId": "low",
       "notes": "Execution explorer: runs bounded checks of runtime behavior where authorized. Validate configured availability at launch; hold affected work without fallback."
@@ -149,7 +149,7 @@
       "id": "axstack-monitor",
       "name": "Axstack monitor",
       "provider": "claude",
-      "model": "claude-opus-5",
+      "model": "claude-opus-5-5",
       "modeId": "bypassPermissions",
       "thinkingOptionId": "medium",
       "notes": "Optional independent read-only observer for a standalone PR watch. Reads GitHub, feedback, and checks, persists event IDs, and wakes the owner only for a new actionable event. Never sends, authors, reviews, replies, or acts as either reusable PR manager. Healthy snapshots stay quiet."
@@ -158,7 +158,7 @@
       "id": "axstack-auditor",
       "name": "Axstack auditor",
       "provider": "codex",
-      "model": "gpt-5.6-luna",
+      "model": "gpt-6-luna",
       "modeId": "full-access",
       "thinkingOptionId": "max",
       "notes": "Read-only end-of-run and checkpoint auditor. Collects scope and outcome evidence with counts and denominators and reports PASS, FAIL, or UNKNOWN without inventing numbers. Never edits, merges, activates, or audits itself."
@@ -167,7 +167,7 @@
       "id": "axstack-debug-investigator-1",
       "name": "Axstack debug investigator 1",
       "provider": "claude",
-      "model": "claude-opus-5",
+      "model": "claude-opus-5-5",
       "modeId": "bypassPermissions",
       "thinkingOptionId": "medium",
       "notes": "Debug investigator seat 1. Dispatched only by axstack-debug at L1 with the shared evidence packet and one distinct brief; never reads another investigator's output. Works in its own disposable worktree at the pinned revision plus the recorded dirty patch; may instrument there for probes; never commits, pushes, publishes, or creates children. Returns one receipt per brief. Independence comes from brief isolation, not model diversity."
@@ -176,7 +176,7 @@
       "id": "axstack-debug-investigator-2",
       "name": "Axstack debug investigator 2",
       "provider": "codex",
-      "model": "gpt-5.6-sol",
+      "model": "gpt-6-sol",
       "modeId": "full-access",
       "thinkingOptionId": "medium",
       "notes": "Debug investigator seat 2. Dispatched only by axstack-debug at L1 with the shared evidence packet and one distinct brief; never reads another investigator's output. Works in its own disposable worktree at the pinned revision plus the recorded dirty patch; may instrument there for probes; never commits, pushes, publishes, or creates children. Returns one receipt per brief. Independence comes from brief isolation, not model diversity."
@@ -194,7 +194,7 @@
       "id": "axstack-debug-investigator-4",
       "name": "Axstack debug investigator 4",
       "provider": "codex",
-      "model": "gpt-5.6-terra",
+      "model": "gpt-6-sol",
       "modeId": "full-access",
       "thinkingOptionId": "low",
       "notes": "Debug investigator seat 4. Dispatched only by axstack-debug at L1 with the shared evidence packet and one distinct brief; never reads another investigator's output. Works in its own disposable worktree at the pinned revision plus the recorded dirty patch; may instrument there for probes; never commits, pushes, publishes, or creates children. Returns one receipt per brief. Independence comes from brief isolation, not model diversity."

package/skills/axstack/references/evidence-archive.md

@@ -59,12 +59,44 @@ state and require exit proof for that exact terminal incarnation. A task
 terminal, manual chat, unexpected terminal, failed close, or uncertain exit
 remains protected.
 
-After receipt and continuity readback, if the only remaining Git dirt is the
-verified archived untracked evidence, compare its current bytes with the
-manifest again and unlink only those exact regular evidence files individually.
-Never remove tracked or unknown files, directories, or any path whose hash now
-differs. Re-read Git and native state; any remaining or uncertain dirt holds
-retirement.
+After receipt and continuity readback, invoke the same installed helper with
+the same identity and complete `--file` set, plus the recorded manifest hash:
+
+```sh
+bun scripts/archive-evidence.js \
+  --source-root <absolute-worktree-or-evidence-root> \
+  --archive-root <absolute-private-archive-root> \
+  --repo <owner/repository> --pr <number> --head <40-character-sha> \
+  --dispatch <exact-dispatch-id> \
+  --file <classified-relative-file> [--file <classified-relative-file> ...] \
+  --operation retire --manifest-hash <recorded-64-character-sha256>
+```
+
+Use the corresponding `--run` and `--task` identity for a non-PR archive. The
+retirement operation independently verifies the immutable private archive, its
+exact identity and complete file set, the caller-recorded manifest hash, the
+exact Git top-level and HEAD, and every remaining source file. It refuses
+tracked, staged, or unclassified dirt; changed bytes; unsafe or overlapping
+roots; symlinks; hard links; and non-regular files. It applies exact unlinks only
+to matching listed files and never removes directories.
+
+Record the retirement receipt's `removed`, `alreadyAbsent`, and `pending` file
+sets. A repeat invocation reconciles already absent files without changing the
+manifest. A partial or failed invocation preserves the archive; resolve its
+exact hold and retry the same operation until `pending` is empty. Never replace
+this operation with a shell loop, broad deletion, force, or a waiver.
+
+Re-read Git and native state after successful retirement. Any remaining or
+uncertain dirt holds worktree removal. Settlement, liveness/no-writer proof,
+useful-work and publication checks, evidence classification, and removal
+authority remain driver decisions; archive or retirement success proves none of
+them.
+
+Before native worktree removal, verify the effective **Archive Script**
+provenance. An unknown hook or a required hook whose provenance is not trusted
+holds removal. Record its native outcome as exactly `unconfigured`, `passed`,
+`failed`, or `unknown`; only `unconfigured` or a trusted `passed` outcome may
+advance, while `failed` and `unknown` preserve the resource.
 
 Only then use the version-matched Orca guide's native worktree cleanup operation
 with the exact workspace identity. Never use shell recursive deletion and never

package/skills/axstack/references/routing.md

@@ -34,10 +34,10 @@ Role IDs:
 
   | Preset | Author | Reviewer (model/effort) |
   | --- | --- | --- |
-  | `mixed` | Codex / Sol (`codex/gpt-5.6-sol`) | `axstack-reviewer-secondary` (`claude/claude-opus-5` medium) |
-  | `mixed` | Claude / Opus (`claude/claude-opus-5`) | `axstack-reviewer-primary` (`codex/gpt-5.6-sol` medium) |
-  | `codex-only` | Codex / Sol (`codex/gpt-5.6-sol`) | `axstack-reviewer-secondary` (`codex/gpt-5.6-terra` xhigh) |
-  | `claude-only` | Claude / Opus (`claude/claude-opus-5`) | `axstack-reviewer-secondary` (`claude/claude-sonnet-5` xhigh) |
+  | `mixed` | Codex / Sol (`codex/gpt-6-sol`) | `axstack-reviewer-secondary` (`claude/claude-opus-5-5` medium) |
+  | `mixed` | Claude / Opus (`claude/claude-opus-5-5`) | `axstack-reviewer-primary` (`codex/gpt-6-sol` medium) |
+  | `codex-only` | Codex / Sol (`codex/gpt-6-sol`) | `axstack-reviewer-secondary` (`codex/gpt-6-luna` xhigh) |
+  | `claude-only` | Claude / Opus (`claude/claude-opus-5-5`) | `axstack-reviewer-secondary` (`claude/claude-sonnet-5` xhigh) |
 - `axstack-advisor-astra` and `axstack-advisor-fable` advise independently
   and author align arena candidates; `axstack-arena-judge-astra` and
   `axstack-arena-judge-fable` judge them. `axstack-auditor` audits;

package/skills/axstack/scripts/archive-evidence.js

@@ -7,6 +7,7 @@ import {
   readFile,
   rename,
   rm,
+  unlink,
   writeFile,
 } from 'node:fs/promises';
 
@@ -68,7 +69,7 @@ function parseArgs(argv) {
     if (!flag?.startsWith('--') || value === undefined) fail(`invalid argument near ${flag ?? '(end)'}`);
     const key = flag.slice(2);
     if (key === 'file') values.files.push(value);
-    else if (['source-root', 'archive-root', 'repo', 'pr', 'run', 'task', 'head', 'dispatch'].includes(key)) {
+    else if (['source-root', 'archive-root', 'repo', 'pr', 'run', 'task', 'head', 'dispatch', 'operation', 'manifest-hash'].includes(key)) {
       if (values[key] !== undefined) fail(`duplicate --${key}`);
       values[key] = value;
     } else fail(`unknown argument: ${flag}`);
@@ -76,6 +77,7 @@ function parseArgs(argv) {
   for (const key of ['source-root', 'archive-root', 'repo', 'head', 'dispatch']) {
     if (!values[key]) fail(`missing --${key}`);
   }
+  values.operation ??= 'archive';
   if (values.files.length === 0) fail('at least one --file is required');
   if (!isAbsolute(values['source-root']) || !isAbsolute(values['archive-root'])) {
     fail('source and archive roots must be absolute');
@@ -95,6 +97,13 @@ function parseArgs(argv) {
   }
   if (!/^[0-9a-f]{40}$/.test(values.head)) fail('head must be an exact 40-character lowercase SHA');
   if (!/^[A-Za-z0-9_-]+$/.test(values.dispatch)) fail('dispatch contains unsafe characters');
+  if (!['archive', 'retire'].includes(values.operation)) fail('operation must be archive or retire');
+  if (values.operation === 'retire' && !/^[0-9a-f]{64}$/.test(values['manifest-hash'] ?? '')) {
+    fail('retirement requires an exact lowercase --manifest-hash');
+  }
+  if (values.operation === 'archive' && values['manifest-hash'] !== undefined) {
+    fail('--manifest-hash applies only to retirement');
+  }
   values.files = [...new Set(values.files)].sort();
   for (const file of values.files) {
     const parts = file.split('/');
@@ -212,6 +221,203 @@ async function verifyArchive(archiveDir, identity, collected) {
   };
 }
 
+function sameJson(left, right) {
+  return JSON.stringify(left) === JSON.stringify(right);
+}
+
+async function verifyRetirementArchive(archiveRoot, archiveDir, identity, files, manifestHash) {
+  const archiveRel = relative(archiveRoot, archiveDir);
+  let current = archiveRoot;
+  for (const part of ['', ...archiveRel.split('/')]) {
+    if (part) current = join(current, part);
+    const st = await assertRealPath(current, 'directory');
+    if ((st.mode & 0o077) !== 0) fail(`archive permissions are not private: ${current}`);
+  }
+
+  const manifestPath = join(archiveDir, 'manifest.json');
+  const manifestStat = await assertRealPath(manifestPath, 'file');
+  if (manifestStat.nlink !== 1) fail(`refusing hard-linked archive file: ${manifestPath}`);
+  if ((manifestStat.mode & 0o077) !== 0) fail(`archive manifest permissions are not private: ${manifestPath}`);
+  const manifestBytes = await readFile(manifestPath);
+  if (sha256(manifestBytes) !== manifestHash) fail('recorded manifest hash mismatch');
+
+  let manifest;
+  try {
+    manifest = JSON.parse(manifestBytes.toString());
+  } catch {
+    fail(`invalid archive manifest: ${manifestPath}`);
+  }
+  if (manifest.version !== 1 || !sameJson(manifest.identity, identity)) {
+    fail(`archive manifest identity mismatch: ${manifestPath}`);
+  }
+  const manifestFiles = Object.keys(manifest.files ?? {}).sort();
+  if (!sameJson(manifestFiles, files)) fail(`archive manifest file set mismatch: ${manifestPath}`);
+  if (!manifestBytes.equals(Buffer.from(JSON.stringify(manifest, null, 2) + '\n'))) {
+    fail(`archive manifest bytes are not canonical: ${manifestPath}`);
+  }
+
+  const filesRoot = join(archiveDir, 'files');
+  const filesRootStat = await assertRealPath(filesRoot, 'directory');
+  if ((filesRootStat.mode & 0o077) !== 0) fail(`archive permissions are not private: ${filesRoot}`);
+  for (const rel of files) {
+    const expected = manifest.files[rel];
+    if (!expected || !/^[0-9a-f]{64}$/.test(expected.sha256) || !Number.isSafeInteger(expected.size) || expected.size < 0) {
+      fail(`invalid archive manifest entry: ${rel}`);
+    }
+    await assertNoSymlinkComponents(filesRoot, rel);
+    const archived = join(filesRoot, rel);
+    const st = await assertRealPath(archived, 'file');
+    if (st.nlink !== 1) fail(`refusing hard-linked archive file: ${archived}`);
+    if ((st.mode & 0o077) !== 0) fail(`archived evidence permissions are not private: ${archived}`);
+    const bytes = await readFile(archived);
+    if (bytes.length !== expected.size || sha256(bytes) !== expected.sha256) {
+      fail(`archived evidence hash mismatch: ${rel}`);
+    }
+  }
+  const actualFiles = await listArchiveFiles(archiveDir);
+  const expectedFiles = ['manifest.json', ...files.map((rel) => `files/${rel}`)].sort();
+  if (!sameJson(actualFiles, expectedFiles)) fail(`archive contains unexpected or missing files: ${archiveDir}`);
+  return { archiveDir, manifestPath, manifestHash, files: files.length, manifest };
+}
+
+function gitOutput(sourceRoot, args) {
+  const result = Bun.spawnSync(['git', '-C', sourceRoot, ...args], { stdout: 'pipe', stderr: 'pipe' });
+  if (result.exitCode !== 0) {
+    fail(`git ${args.join(' ')} failed: ${result.stderr.toString().trim() || `exit ${result.exitCode}`}`);
+  }
+  return result.stdout.toString();
+}
+
+async function inspectSource(sourceRoot, rel) {
+  let current = sourceRoot;
+  const parts = rel.split('/');
+  for (let index = 0; index < parts.length; index += 1) {
+    current = join(current, parts[index]);
+    const st = await lstat(current, { bigint: true }).catch((err) => {
+      if (err?.code === 'ENOENT') return null;
+      throw err;
+    });
+    if (!st) return null;
+    if (st.isSymbolicLink()) fail(`refusing symlink: ${current}`);
+    if (index < parts.length - 1 && !st.isDirectory()) fail(`not a directory: ${current}`);
+    if (index === parts.length - 1) {
+      if (!st.isFile()) fail(`not a regular file: ${current}`);
+      if (st.nlink !== 1n) fail(`refusing hard-linked source file: ${current}`);
+      const bytes = await readFile(current);
+      const after = await lstat(current, { bigint: true }).catch((err) => {
+        if (err?.code === 'ENOENT') fail(`source changed during verification: ${rel}`);
+        throw err;
+      });
+      for (const key of ['dev', 'ino', 'size', 'mtimeNs', 'nlink']) {
+        if (st[key] !== after[key]) fail(`source changed during verification: ${rel}`);
+      }
+      return {
+        bytes,
+        sha256: sha256(bytes),
+        size: bytes.length,
+        fingerprint: Object.fromEntries(
+          ['dev', 'ino', 'size', 'mtimeNs', 'ctimeNs', 'nlink'].map((key) => [key, after[key]]),
+        ),
+      };
+    }
+  }
+}
+
+function sameSource(left, right) {
+  return left.size === right.size &&
+    left.sha256 === right.sha256 &&
+    Object.keys(left.fingerprint).every((key) => left.fingerprint[key] === right.fingerprint[key]);
+}
+
+async function verifyRetirementSource(sourceRoot, head, files, manifest) {
+  await assertSafeAncestors(sourceRoot);
+  await assertRealPath(sourceRoot, 'directory');
+  const topLevel = resolve(gitOutput(sourceRoot, ['rev-parse', '--show-toplevel']).trim());
+  if (topLevel !== sourceRoot) fail(`source root is not the exact Git top-level: ${sourceRoot}`);
+  const actualHead = gitOutput(sourceRoot, ['rev-parse', 'HEAD']).trim();
+  if (actualHead !== head) fail(`Git HEAD mismatch: expected ${head}, found ${actualHead}`);
+
+  const states = {};
+  for (const rel of files) {
+    const state = await inspectSource(sourceRoot, rel);
+    if (state && (state.size !== manifest.files[rel].size || state.sha256 !== manifest.files[rel].sha256)) {
+      fail(`source evidence hash mismatch: ${rel}`);
+    }
+    states[rel] = state;
+  }
+
+  const expectedUntracked = new Set(files.filter((rel) => states[rel]));
+  const statusEntries = gitOutput(sourceRoot, [
+    'status', '--porcelain=v1', '-z', '--untracked-files=all', '--ignored=no',
+  ]).split('\0').filter(Boolean);
+  for (const entry of statusEntries) {
+    const code = entry.slice(0, 2);
+    const rel = entry.slice(3);
+    if (code !== '??' || !expectedUntracked.delete(rel)) {
+      fail(`tracked, staged, or unclassified Git dirt: ${rel || '(unknown)'}`);
+    }
+  }
+  if (expectedUntracked.size > 0) {
+    fail(`source evidence is not classified as untracked: ${[...expectedUntracked].sort()[0]}`);
+  }
+  return states;
+}
+
+async function retireEvidence(args, sourceRoot, archiveRoot, archiveDir, identity) {
+  const archive = await verifyRetirementArchive(
+    archiveRoot, archiveDir, identity, args.files, args['manifest-hash'],
+  );
+  const states = await verifyRetirementSource(sourceRoot, args.head, args.files, archive.manifest);
+  const removed = [];
+  const alreadyAbsent = args.files.filter((rel) => !states[rel]);
+
+  for (const rel of args.files) {
+    if (!states[rel]) continue;
+    const current = await inspectSource(sourceRoot, rel);
+    if (!current || !sameSource(current, states[rel])) fail(`source changed before retirement: ${rel}`);
+  }
+
+  for (let index = 0; index < args.files.length; index += 1) {
+    const rel = args.files[index];
+    if (!states[rel]) continue;
+    const current = await inspectSource(sourceRoot, rel);
+    if (!current || !sameSource(current, states[rel])) {
+      fail(`source changed before retirement: ${rel}`);
+    }
+    try {
+      await unlink(join(sourceRoot, rel));
+    } catch (err) {
+      const pending = args.files.slice(index).filter((file) => states[file]);
+      console.log(JSON.stringify({
+        status: 'partial',
+        archiveDir: archive.archiveDir,
+        manifestPath: archive.manifestPath,
+        manifestHash: archive.manifestHash,
+        files: archive.files,
+        removed,
+        alreadyAbsent,
+        pending,
+      }));
+      fail(`retirement stopped at ${rel}: ${err.message}`);
+    }
+    if (await lstat(join(sourceRoot, rel)).catch((err) => err?.code === 'ENOENT' ? null : Promise.reject(err))) {
+      fail(`source still exists after unlink: ${rel}`);
+    }
+    removed.push(rel);
+  }
+
+  console.log(JSON.stringify({
+    status: 'retired',
+    archiveDir: archive.archiveDir,
+    manifestPath: archive.manifestPath,
+    manifestHash: archive.manifestHash,
+    files: archive.files,
+    removed,
+    alreadyAbsent,
+    pending: [],
+  }));
+}
+
 async function listArchiveFiles(root, prefix = '') {
   const files = [];
   const entries = await readdir(join(root, prefix), { withFileTypes: true });
@@ -238,7 +444,6 @@ async function main() {
     archiveToSource === '' || (!archiveToSource.startsWith('..') && !isAbsolute(archiveToSource))
   ) fail('source and archive roots must not contain each other');
 
-  const collected = await collectSource(sourceRoot, args.files);
   const identity = args.pr
     ? { repo: args.repo, pr: Number(args.pr), head: args.head, dispatch: args.dispatch }
     : { repo: args.repo, run: args.run, task: args.task, head: args.head, dispatch: args.dispatch };
@@ -249,6 +454,12 @@ async function main() {
   const archiveDir = join(archiveRoot, repoSlug, ...identityParts, args.head, args.dispatch);
 
   await assertSafeAncestors(archiveRoot);
+  if (args.operation === 'retire') {
+    await retireEvidence(args, sourceRoot, archiveRoot, archiveDir, identity);
+    return;
+  }
+
+  const collected = await collectSource(sourceRoot, args.files);
   let current = archiveRoot;
   const generated = [repoSlug, ...identityParts, args.head];
   await ensurePrivateDir(current);

package/skills/axstack-audit/SKILL.md

@@ -25,7 +25,7 @@ immediately before an actual auditor profile or session dispatch. Ordinary
 audit reading and record writing do not load it, and the auditor never
 dispatches.
 
-Core owns the `axstack-auditor` profile (codex/gpt-5.6-luna max) and its
+Core owns the `axstack-auditor` profile (codex/gpt-6-luna max) and its
 invocation. This skill governs what that auditor reads, measures, and proposes.
 The user-chosen improvement mode is a tested, independently reviewed PR that a
 human merges.

package/skills/axstack-cleanup/SKILL.md

@@ -81,9 +81,15 @@ Treat these operations as separate decisions and receipts:
 3. **Worktree removal.** Re-read Git status, branch/upstream divergence,
    unpushed commits, forge merge/publication state, children, terminals, and
    archived evidence immediately before the native exact-workspace removal.
-   Account for branch-deletion side effects explicitly, then re-list both native
-   workspaces and Git refs. A failed hook or uncertain response preserves the
-   resource; never force or substitute shell deletion.
+   When archived evidence is the last dirt, use the evidence archive helper's
+   manifest-bound retirement operation and require an empty pending set; never
+   unlink through prose or a shell loop. Verify the effective Archive Script
+   provenance before native removal: an unknown or required-but-untrusted hook
+   holds. Record its native outcome as `unconfigured`, `passed`, `failed`, or
+   `unknown`; only `unconfigured` or trusted `passed` may advance. Account for
+   branch-deletion side effects explicitly, then re-list both native workspaces
+   and Git refs. A failed or unknown hook outcome or uncertain response preserves
+   the resource; never force or substitute shell deletion.
 4. **Chat archival.** Attempt it only if the version-matched runtime guide
    advertises a distinct supported operation and the scoped chat is eligible.
    Otherwise record chat archival as unsupported. Process exit, worker release,

package/skills/axstack-explain/SKILL.md

@@ -48,12 +48,24 @@ immediately before an actual profile dispatch.
 
 ## 2. Choose proportional output
 
-1. For a simple request, answer concisely in the current chat. Use a compact
+1. Keep the primary reader-facing explanation to a maximum of 700 words in
+   chat, HTML, and every other requested format. Preserve in that primary view
+   the answer or purpose, key rationale, meaningful alternatives, main data or
+   operational boundary, status and uncertainty, and live reader questions
+   with evidence-supported answers, labeling any question the inspected
+   evidence does not answer as **unknown** or **open** rather than inventing an
+   answer. Count the primary artifact's reader-visible words, including
+   headings, table text, and diagram or figure labels and captions. An appendix
+   or collapsible content in the same artifact counts toward the 700-word
+   maximum. If the draft is longer, compress repetition first and move only
+   supporting detail to a separate linked ticket or appendix. Essential answers
+   must not be hidden behind links, and evidence must not be silently discarded.
+2. For a simple request, answer concisely in the current chat. Use a compact
    diagram when useful. This needs no mandatory agent or intermediate artifact.
-2. For a complex visual, use the configured `axstack-explainer` role to create
+3. For a complex visual, use the configured `axstack-explainer` role to create
    self-contained HTML, or use the requested artifact format. An explicit user
    theme wins; otherwise use the dark default.
-3. Profile IDs are presets, not availability proof. Before dispatch, follow the
+4. Profile IDs are presets, not availability proof. Before dispatch, follow the
    launch sequence and preserve the configured model, mode, and effort. Report
    an unavailable route; never substitute a model.
 

package/skills/axstack-review/SKILL.md

@@ -116,10 +116,10 @@ owns the event and settles after its skill-owned reviewers settle.
 
      | Preset | Actual author provider/model | Reviewer role (configured model/effort) |
      | --- | --- | --- |
-     | `mixed` | Codex / Sol (`codex/gpt-5.6-sol`) | `axstack-reviewer-secondary` (`claude/claude-opus-5` medium) |
-     | `mixed` | Claude / Opus (`claude/claude-opus-5`) | `axstack-reviewer-primary` (`codex/gpt-5.6-sol` medium) |
-     | `codex-only` | Codex / Sol (`codex/gpt-5.6-sol`) | `axstack-reviewer-secondary` (`codex/gpt-5.6-terra` xhigh) |
-     | `claude-only` | Claude / Opus (`claude/claude-opus-5`) | `axstack-reviewer-secondary` (`claude/claude-sonnet-5` xhigh) |
+     | `mixed` | Codex / Sol (`codex/gpt-6-sol`) | `axstack-reviewer-secondary` (`claude/claude-opus-5-5` medium) |
+     | `mixed` | Claude / Opus (`claude/claude-opus-5-5`) | `axstack-reviewer-primary` (`codex/gpt-6-sol` medium) |
+     | `codex-only` | Codex / Sol (`codex/gpt-6-sol`) | `axstack-reviewer-secondary` (`codex/gpt-6-luna` xhigh) |
+     | `claude-only` | Claude / Opus (`claude/claude-opus-5-5`) | `axstack-reviewer-secondary` (`claude/claude-sonnet-5` xhigh) |
 
      Provenance is matched on provider/model ID; record effort, but never use
      effort to create a mapping. Any other author provenance for the

package/src/roles.js

@@ -6,14 +6,14 @@ const PROVIDER_BOUNDS = Object.freeze({
 
 const AUTHORED_ROUTES = Object.freeze({
   mixed: {
-    'codex/gpt-5.6-sol': ['axstack-reviewer-secondary', 'claude/claude-opus-5', 'medium'],
-    'claude/claude-opus-5': ['axstack-reviewer-primary', 'codex/gpt-5.6-sol', 'medium'],
+    'codex/gpt-6-sol': ['axstack-reviewer-secondary', 'claude/claude-opus-5-5', 'medium'],
+    'claude/claude-opus-5-5': ['axstack-reviewer-primary', 'codex/gpt-6-sol', 'medium'],
   },
   'codex-only': {
-    'codex/gpt-5.6-sol': ['axstack-reviewer-secondary', 'codex/gpt-5.6-terra', 'xhigh'],
+    'codex/gpt-6-sol': ['axstack-reviewer-secondary', 'codex/gpt-6-luna', 'xhigh'],
   },
   'claude-only': {
-    'claude/claude-opus-5': ['axstack-reviewer-secondary', 'claude/claude-sonnet-5', 'xhigh'],
+    'claude/claude-opus-5-5': ['axstack-reviewer-secondary', 'claude/claude-sonnet-5', 'xhigh'],
   },
 });