dsh-aris-panel 0.1.0

package/templates/PAPER_PLAN_TEMPLATE.md

@@ -0,0 +1,47 @@
+# Paper Plan
+
+> **Template for Workflow 3 — skip planning phase.** Fill in, then run `/paper-writing "PAPER_PLAN.md"`.
+
+## Metadata
+- **Title**: [Title]
+- **Venue**: [ICLR / NeurIPS / ICML]
+- **One-sentence contribution**: [Core takeaway]
+
+## Claims-Evidence Matrix
+| # | Claim | Evidence | Section |
+|---|-------|----------|---------|
+| C1 | [Main claim] | [Table 1, Exp A] | §3 |
+| C2 | [Supporting] | [Figure 2] | §4 |
+
+## Section Plan
+
+### 1. Introduction (~1.5 pages)
+- **What**: [contribution]
+- **Why**: [importance]
+- **How**: [approach]
+- **Result**: [strongest number]
+
+### 2. Related Work (~1 page)
+- [Group 1]: [papers, gap]
+- [Group 2]: [papers, gap]
+
+### 3. Method (~2 pages)
+- [Problem formulation]
+- [Proposed approach]
+
+### 4. Experiments (~3 pages)
+- [Setup, main results, ablation]
+
+### 5. Conclusion (~0.5 pages)
+- [Summary, limitations, future]
+
+## Figure Plan
+| # | Type | Description | Auto? |
+|---|------|-------------|:-----:|
+| Fig 1 | Architecture | Method overview | illustration |
+| Fig 2 | Bar chart | Main results | matplotlib |
+| Table 1 | Comparison | SOTA | LaTeX |
+
+## Key References
+1. [Author et al., "Title", Venue Year]
+2. [...]

package/templates/PATENT_CLAIMS_TEMPLATE.md

@@ -0,0 +1,78 @@
+# Patent Claims Worksheet
+
+> **Template for pre-drafting claims structure.** Fill in before running `/claims-drafting`, or use as a reference during claim drafting.
+
+## Invention Summary
+[One-sentence description of the core inventive concept.]
+
+## Independent Claims Plan
+
+### Claim 1: Method Claim (broadest)
+- **Preamble**: A method for [purpose], comprising:
+- **Known features** (from prior art):
+  1. [feature A]
+  2. [feature B]
+- **Inventive features** (novel):
+  1. [feature C]
+  2. [feature D]
+- **Transition**: comprising / characterised in that (per jurisdiction)
+
+### Claim X: System/Apparatus Claim (mirrors Claim 1)
+- **Preamble**: A system for [purpose], comprising:
+- **Structural components mapping to method steps**:
+  1. [component corresponding to step 1]
+  2. [component corresponding to step 2]
+  ...
+
+### Claim Y: Computer-Readable Medium Claim (US only, optional)
+- Maps to method claim 1
+
+## Dependent Claims Plan
+
+| Claim # | Depends On | Adds Feature | Fallback Value |
+|---------|-----------|-------------|----------------|
+| 2 | 1 | [specific limitation of feature C] | Medium |
+| 3 | 1 | [specific limitation of feature D] | Medium |
+| 4 | 2 | [additional narrowing] | Low |
+| 5 | 1 or 2 | [alternative implementation of C] | High |
+
+## Claims Hierarchy (dependency tree)
+
+```
+1. (Independent — method, broadest)
+├── 2. (narrows feature C)
+│   └── 4. (narrows further)
+├── 3. (narrows feature D)
+├── 5. (alternative to 2 or 3)
+│
+X. (Independent — system, mirrors 1)
+├── X+1. (narrows component)
+├── X+2. (narrows component)
+│
+Y. (Independent — medium, US only, mirrors 1)
+```
+
+## Prior Art Avoidance Matrix
+
+| Claim Element | Prior Art Reference | Avoidance Strategy |
+|---------------|--------------------|--------------------|
+| Feature C | [Ref X] | Ref X uses different approach... |
+| Feature D | [Ref Y] | Ref Y doesn't combine C+D... |
+| Combination C+D | [Ref X + Y] | No motivation to combine... |
+
+## Jurisdiction-Specific Formatting Notes
+
+### CN (Chinese)
+- Use "其特征在于" to separate known and inventive features
+- Method: "一种...的方法，包括：...；其特征在于，还包括：..."
+- Apparatus: "一种...的装置，包括：...；其特征在于，还包括：..."
+
+### US
+- Use "comprising" (open transition)
+- Method: "A method for [purpose], comprising: [step A]; [step B];..."
+- System: "A system for [purpose], comprising: [component A]; [component B];..."
+
+### EP
+- Two-part form MANDATORY (Rule 43(1) EPC)
+- "A method for [purpose], comprising [known features], characterised in that [inventive features]."
+- "A system for [purpose], comprising [known components], characterised in that [inventive components]."

package/templates/PATENT_SPECIFICATION_TEMPLATE.md

@@ -0,0 +1,68 @@
+# Patent Specification Skeleton
+
+> **Template for `/specification-writing`.** Provides pre-structured sections. The skill fills in content based on INVENTION_DISCLOSURE.md and CLAIMS.md.
+
+## Metadata
+- **Title**: [发明名称 / Title of Invention]
+- **Jurisdiction**: [CN / US / EP / ALL]
+- **Patent Type**: [Invention / Utility Model]
+- **Language**: [Chinese / English]
+
+## 1. Technical Field (技术领域)
+[1-2 paragraphs identifying the technical domain]
+
+## 2. Background Technology (背景技术)
+### 2.1 Existing Approaches
+[Describe the closest prior art — patents, papers, or products]
+
+### 2.2 Deficiencies
+[Identify specific technical deficiencies of existing approaches]
+
+## 3. Summary of Invention (发明内容)
+
+### 3.1 Technical Problem
+[State the problem derived from background deficiencies]
+
+### 3.2 Technical Solution
+[Describe the invention's solution, matching the scope of the claims]
+
+### 3.3 Advantages
+[Measurable/quantifiable benefits over prior art]
+
+## 4. Brief Description of Drawings (附图说明)
+
+| Figure | Description |
+|--------|-------------|
+| FIG. 1 | [one-sentence description] |
+| FIG. 2 | [one-sentence description] |
+| FIG. 3 | [one-sentence description] |
+
+## 5. Detailed Description of Embodiments (具体实施方式)
+
+### Embodiment 1: [Title]
+[Detailed description with reference numerals. How to make and use the invention.]
+
+### Embodiment 2: [Title — optional]
+[Alternative embodiment supporting broader claim interpretation]
+
+## 6. Abstract (摘要)
+[150 words (US) / 300 words (CN). Summary of technical solution and key advantages.]
+
+## Reference Numeral Index
+
+| Numeral | Component | Figure(s) |
+|---------|-----------|-----------|
+| 100 | [system] | Fig. 1 |
+| 102 | [processor] | Fig. 1 |
+| 104 | [memory] | Fig. 1 |
+| 200 | [process] | Fig. 2 |
+| 202 | [step A] | Fig. 2 |
+
+## Claim-to-Specification Support Map
+
+| Claim Element | Specification Paragraph(s) | Figure(s) |
+|---------------|---------------------------|-----------|
+| [feature A] | ¶X | Fig. 1 |
+| [feature B] | ¶X-Y | Fig. 1, 2 |
+| [feature C] | ¶Z | Fig. 2 |
+| [feature D] | ¶Z-A | Fig. 2, 3 |

package/templates/README.md

@@ -0,0 +1,57 @@
+# Templates
+
+Ready-to-use templates for each ARIS workflow. Copy, fill in your content, and run the corresponding skill.
+
+### Workflow Input Templates
+
+| Template | For Workflow | What to do |
+|----------|-------------|------------|
+| [RESEARCH_BRIEF_TEMPLATE.md](RESEARCH_BRIEF_TEMPLATE.md) | Workflow 1 | Detailed research direction as document input |
+| [RESEARCH_CONTRACT_TEMPLATE.md](RESEARCH_CONTRACT_TEMPLATE.md) | Workflow 1 | Define problem boundaries, non-goals, timeline before starting |
+| [EXPERIMENT_PLAN_TEMPLATE.md](EXPERIMENT_PLAN_TEMPLATE.md) | Workflow 1.5 | Claim-driven experiment roadmap with run order and budgets |
+| [NARRATIVE_REPORT_TEMPLATE.md](NARRATIVE_REPORT_TEMPLATE.md) | Workflow 3 | Research narrative with claims, experiments, results |
+| [PAPER_PLAN_TEMPLATE.md](PAPER_PLAN_TEMPLATE.md) | Workflow 3 | Pre-made outline to skip planning phase |
+| [CLAUDE_MD_TEMPLATE.md](CLAUDE_MD_TEMPLATE.md) | All Workflows | Project dashboard with Pipeline Status — create once per project |
+| [MANIFEST_TEMPLATE.md](MANIFEST_TEMPLATE.md) | All Workflows | Output tracking manifest — auto-maintained by skills |
+
+### Chinese Templates (中文模板)
+
+| Template | For Workflow | What to do |
+|----------|-------------|------------|
+| [RESEARCH_BRIEF_TEMPLATE_CN.md](RESEARCH_BRIEF_TEMPLATE_CN.md) | Workflow 1 | 研究简报中文模板 |
+| [IDEA_CANDIDATES_TEMPLATE_CN.md](IDEA_CANDIDATES_TEMPLATE_CN.md) | Workflow 1 | Idea 候选池中文模板 |
+| [EXPERIMENT_PLAN_TEMPLATE_CN.md](EXPERIMENT_PLAN_TEMPLATE_CN.md) | Workflow 1.5 | 实验计划中文模板 |
+
+### Patent Templates (`/patent-pipeline`)
+
+| Template | For Workflow | What to do |
+|----------|-------------|------------|
+| [INVENTION_BRIEF_TEMPLATE.md](INVENTION_BRIEF_TEMPLATE.md) | Patent Pipeline | Invention disclosure with technical problem, solution, advantages, figures |
+| [PATENT_CLAIMS_TEMPLATE.md](PATENT_CLAIMS_TEMPLATE.md) | `/claims-drafting` | Claims hierarchy worksheet with examples for CN/US/EP |
+| [PATENT_SPECIFICATION_TEMPLATE.md](PATENT_SPECIFICATION_TEMPLATE.md) | `/specification-writing` | Skeleton specification with all required sections |
+
+### Compact Mode Templates (`— compact: true`)
+
+| Template | Written by | Purpose |
+|----------|-----------|---------|
+| [IDEA_CANDIDATES_TEMPLATE.md](IDEA_CANDIDATES_TEMPLATE.md) | `/idea-discovery` | Top 3-5 surviving ideas (lean, not full 12-idea report) |
+| [EXPERIMENT_LOG_TEMPLATE.md](EXPERIMENT_LOG_TEMPLATE.md) | `/experiment-bridge` | Structured experiment record (results + reproduction commands) |
+| [FINDINGS_TEMPLATE.md](FINDINGS_TEMPLATE.md) | `/auto-review-loop` | One-line-per-finding discovery log (anomalies, decisions) |
+
+## Usage
+
+### Research Pipeline
+
+```bash
+cp templates/EXPERIMENT_PLAN_TEMPLATE.md refine-logs/EXPERIMENT_PLAN.md
+# Edit with your content, then:
+/experiment-bridge
+```
+
+### Patent Pipeline
+
+```bash
+cp templates/INVENTION_BRIEF_TEMPLATE.md patent/INVENTION_BRIEF.md
+# Edit with your invention details, then:
+/patent-pipeline "patent/INVENTION_BRIEF.md -- CN"
+```

package/templates/RESEARCH_BRIEF_TEMPLATE.md

@@ -0,0 +1,35 @@
+# Research Brief
+
+> **Template for document-based input to `/idea-discovery` or `/research-pipeline`.** Provide detailed context instead of a one-line prompt.
+>
+> **Length guideline:** keep the core brief to **1–2 pages (~4–8k chars)**. The research-wiki `query_pack.md` that downstream skills consume is hard-capped at 8,000 chars — an oversized brief crowds out gaps / failed-ideas / paper context. Put long material (full notes, logs, draft text) in separate files and reference them by path.
+
+## Problem Statement
+[2-3 paragraphs: What is broken/missing in current approaches?]
+
+## Background
+- **Field**: [e.g., NLP, CV, RL]
+- **Sub-area**: [e.g., discrete diffusion models]
+- **Key papers I've read**: [3-5 papers]
+- **What I already tried**: [Prior experiments]
+- **What didn't work**: [Failed attempts and why]
+
+## Constraints
+- **Compute**: [e.g., 8x RTX 3090, 100 GPU-hours]
+- **Timeline**: [e.g., 3 months to submission]
+- **Target venue**: [e.g., NeurIPS 2026]
+
+## What I'm Looking For
+- [ ] New research direction from scratch
+- [ ] Improvement on existing method: [which?]
+- [ ] Diagnostic study / analysis paper
+- [ ] Other: [describe]
+
+## Domain Knowledge
+[Intuitions, hypotheses, domain-specific insights]
+
+## Non-Goals
+[What you do NOT want to work on]
+
+## Existing Results (if any)
+[Preliminary results, tables, observations — even negative results help]

package/templates/RESEARCH_BRIEF_TEMPLATE_CN.md

@@ -0,0 +1,41 @@
+# 研究简报
+
+> **用于 `/idea-discovery` 或 `/research-pipeline` 的文档化输入模板。** 提供详细上下文，替代一行式提示。
+>
+> **长度建议**：核心简报控制在 **1–2 页（约 4–8k 字符）**。下游消费的 research-wiki `query_pack.md` 有 8,000 字符硬上限——简报过长会挤占 gaps / 失败 idea / 论文上下文的空间。长材料（完整笔记、日志、草稿）放单独文件，按路径引用。
+
+## 问题陈述
+[2-3 段：当前方法存在什么问题/缺失？]
+
+## 背景
+
+- **领域**: [如：NLP、CV、RL]
+- **子方向**: [如：离散扩散模型]
+- **已读关键论文**: [3-5 篇]
+- **已尝试的方法**: [之前的实验]
+- **失败经验**: [哪些方法没有效果，为什么]
+
+## 约束条件
+
+- **算力**: [如：8x RTX 3090，100 GPU 小时]
+- **时间线**: [如：3 个月内投稿]
+- **目标会议/期刊**: [如：NeurIPS 2026]
+
+## 期望方向
+
+- [ ] 从零开始探索新研究方向
+- [ ] 改进现有方法：[哪个？]
+- [ ] 诊断性研究 / 分析型论文
+- [ ] 其他：[描述]
+
+## 领域知识
+
+[直觉、假设、领域特定的洞察]
+
+## 非目标
+
+[明确不想做的事情]
+
+## 已有结果（如有）
+
+[初步结果、表格、观察 — 即使是负面结果也有帮助]

package/templates/RESEARCH_CONTRACT_TEMPLATE.md

@@ -0,0 +1,60 @@
+# Research Contract: [Idea Name]
+
+> **A focused working document for the currently selected idea.** Created when an idea is chosen from `IDEA_REPORT.md`, updated throughout implementation and training. This is what the LLM reads on session recovery — not the full IDEA_REPORT with all 8-12 candidates.
+>
+> **Why this file exists:** After brainstorming, `IDEA_REPORT.md` contains many candidate ideas. Keeping all of them in context pollutes the LLM's working memory and degrades output quality. This contract extracts *only the active idea* into a standalone document, so new sessions and post-compaction recovery load focused context instead of the entire idea pool.
+
+## Selected Idea
+
+- **Description**: [One-paragraph summary of the idea]
+- **Source**: IDEA_REPORT.md, Idea #N
+- **Selection rationale**: [Why this idea over others — pilot results, novelty score, feasibility]
+
+## Core Claims
+
+1. [Main claim — what your method achieves]
+2. [Supporting claim — why it works / when it works best]
+3. [Optional: scope/limitation claim]
+
+## Method Summary
+
+[2-3 paragraphs: How the method works. Enough detail that a new session can understand the approach without reading the full codebase.]
+
+## Experiment Design
+
+- **Datasets**: [Which datasets, which splits]
+- **Baselines**: [What you compare against]
+- **Metrics**: [Primary and secondary metrics]
+- **Key hyperparameters**: [The ones that matter most]
+- **Compute budget**: [GPU hours, hardware]
+
+## Baselines
+
+| Method | Dataset | Metric | Score | Source |
+|--------|---------|--------|-------|--------|
+| [Baseline A] | [Dataset] | [Metric] | [Number] | [Paper / reproduced] |
+| [Baseline B] | [Dataset] | [Metric] | [Number] | [Paper / reproduced] |
+
+## Current Results
+
+> Updated as experiments complete. Start empty, fill in as you go.
+
+| Method | Dataset | Metric | Score | Notes |
+|--------|---------|--------|-------|-------|
+| [Your method] | [Dataset] | [Metric] | [Number] | [e.g., "3 seeds, mean±std"] |
+
+## Key Decisions
+
+- [Decision 1: Why approach X over Y — with reasoning]
+- [Decision 2: Why this hyperparameter / architecture choice]
+- [Known limitations / risks and how you plan to handle them]
+
+## Status
+
+- [ ] Idea selected
+- [ ] Baseline reproduced
+- [ ] Main method implemented
+- [ ] Representative dataset results
+- [ ] Full dataset results
+- [ ] Ablation studies
+- [ ] Paper draft

package/templates/claude-hooks/corpus_write_guard.json

@@ -0,0 +1,16 @@
+{
+  "_comment": "Merge this into your project's .claude/settings.json. Blocks the COMMON Bash shell-write patterns to the skill corpus (a regex blacklist, NOT a sandbox — a deliberately obfuscated write is not all caught; detection via provenance integrity is the backstop, not this regex). With a read-only producer (meta-optimize / corpus-audit) holding Bash but no Write/Edit, this removes the casual Bash-write path; it does not gate the Write/Edit tools. Adjust the path to corpus_write_guard.py for your checkout (or resolve via the canonical chain).",
+  "hooks": {
+    "PreToolUse": [
+      {
+        "matcher": "Bash",
+        "hooks": [
+          {
+            "type": "command",
+            "command": "python3 \"${CLAUDE_PROJECT_DIR}/templates/claude-hooks/corpus_write_guard.py\""
+          }
+        ]
+      }
+    ]
+  }
+}

package/templates/claude-hooks/corpus_write_guard.py

@@ -0,0 +1,85 @@
+#!/usr/bin/env python3
+"""PreToolUse guard: deny Bash writes to the ARIS skill corpus.
+
+Why: a read-only PRODUCER skill (meta-optimize, corpus-audit) is given no `Write`/
+`Edit` tool so it cannot mutate the corpus with the frictionless mutators. But a skill
+that retains *any* `Bash` access can still write a corpus file via shell redirection
+(`cat > skills/x`, `sed -i ... skills/x`, `tee skills/x`, `cp/mv ... skills/x`) — which
+no `allowed-tools` allowlist can block, because redirection is appended after the matched
+command prefix. The only place to enforce "Bash may not write the corpus" is the harness:
+this PreToolUse hook inspects the actual command and DENIES corpus-write shell ops.
+
+Invariant this establishes (with the producer holding Bash-but-no-Write/Edit):
+    corpus mutation MUST go through the Write/Edit tools (reviewable, attributable) —
+    never an opaque Bash redirection. So a producer with no Write/Edit + this hook
+    active is STRUCTURALLY unable to mutate the corpus by any path.
+
+Scope: this guard is GLOBAL (not skill-scoped — hooks have no reliable skill context).
+It blocks Bash *corpus writes* for every skill; legitimate corpus edits (yours, and
+/meta-apply's landing) use the Write/Edit tools and are unaffected. Blocking opaque
+shell writes to the corpus is good hygiene regardless of who runs.
+
+Threat model: closes the ACCIDENTAL/casual self-acquittal path (an honest producer
+slipping into `cat > skills/...`). A determined adversary can still obfuscate a write
+(`p=skills/x; printf ... >"$p"`); defense-in-depth against that is the provenance
+integrity check (an un-stamped / stale-hash corpus change is detectable before push),
+not this regex. This hook is the cheap, harness-level prevention layer, not a sandbox.
+
+Install: add to .claude/settings.json hooks.PreToolUse (matcher "Bash"); see
+templates/claude-hooks/corpus_write_guard.json. Exit 2 = block + reason to the model.
+"""
+
+from __future__ import annotations
+
+import json
+import re
+import sys
+
+# Corpus = the parts of the repo a self-modification producer must never write via Bash.
+_CORPUS = r"(?:\./)?(?:skills|shared-references|tools|templates|plugins)/"
+
+# Shell ops that WRITE a path. Each pattern is "write-operator ... targeting a corpus path".
+_WRITE_OPS = [
+    re.compile(r">>?\s*\"?'?" + _CORPUS),                      # > corpus / >> corpus
+    re.compile(r"\btee\s+(?:-a\s+)?\"?'?" + _CORPUS),           # tee [-a] corpus
+    re.compile(r"\bsed\s+(?:-[a-zA-Z]*\s+)*-i\b[^|;&]*" + _CORPUS),  # sed -i ... corpus
+    re.compile(r"\bdd\b[^|;&]*\bof=\"?'?" + _CORPUS),           # dd of=corpus
+    re.compile(r"\b(?:cp|mv|install|rsync|ln)\b[^|;&]*\s\"?'?" + _CORPUS),  # cp/mv/... corpus (as dest)
+    re.compile(r"\btruncate\b[^|;&]*" + _CORPUS),
+    re.compile(r"\btouch\s+[^|;&]*" + _CORPUS),                  # touch corpus (create)
+    re.compile(r"\.write_(?:text|bytes)\b[^|;&]*" + _CORPUS),    # Path(...).write_text into corpus
+    re.compile(_CORPUS + r"[^'\"]*['\"]\s*\)\s*\.write_(?:text|bytes)"),  # Path('corpus..').write_text
+    re.compile(r"\b(?:python3?|perl|ruby|node)\b[^|;&]*open\([^)]*" + _CORPUS),  # open('skills/..','w')
+]
+
+
+def violates(command: str) -> str | None:
+    for rx in _WRITE_OPS:
+        m = rx.search(command)
+        if m:
+            return command[m.start():m.start() + 80]
+    return None
+
+
+def main() -> int:
+    try:
+        data = json.load(sys.stdin)
+    except Exception:
+        return 0  # malformed input → don't block (fail open on parse, not on match)
+    if data.get("tool_name") != "Bash":
+        return 0
+    command = (data.get("tool_input") or {}).get("command", "") or ""
+    hit = violates(command)
+    if hit:
+        sys.stderr.write(
+            "BLOCKED by corpus_write_guard: Bash may not WRITE the skill corpus "
+            f"(matched: {hit!r}). Corpus mutation must go through the Write/Edit tools "
+            "(reviewable, attributable). A read-only producer (meta-optimize / corpus-"
+            "audit) stages patches to .aris/meta/ and hands off to /meta-apply, which "
+            "lands them with the Write/Edit tools after the cross-model jury + human gate.\n")
+        return 2  # PreToolUse: block the call, show stderr to the model
+    return 0
+
+
+if __name__ == "__main__":
+    raise SystemExit(main())

package/templates/claude-hooks/meta_logging.json

@@ -0,0 +1,74 @@
+{
+  "$comment": "ARIS Meta-Optimize: Passive event logging for harness optimization. Copy to .claude/settings.json or merge into existing settings.",
+  "hooks": {
+    "PostToolUse": [
+      {
+        "matcher": "",
+        "hooks": [
+          {
+            "type": "command",
+            "command": "bash \"$CLAUDE_PROJECT_DIR/tools/meta_opt/log_event.sh\"",
+            "timeout": 5,
+            "async": true
+          }
+        ]
+      }
+    ],
+    "PostToolUseFailure": [
+      {
+        "matcher": "",
+        "hooks": [
+          {
+            "type": "command",
+            "command": "bash \"$CLAUDE_PROJECT_DIR/tools/meta_opt/log_event.sh\"",
+            "timeout": 5,
+            "async": true
+          }
+        ]
+      }
+    ],
+    "UserPromptSubmit": [
+      {
+        "hooks": [
+          {
+            "type": "command",
+            "command": "bash \"$CLAUDE_PROJECT_DIR/tools/meta_opt/log_event.sh\"",
+            "timeout": 5,
+            "async": true
+          }
+        ]
+      }
+    ],
+    "SessionStart": [
+      {
+        "matcher": "",
+        "hooks": [
+          {
+            "type": "command",
+            "command": "bash \"$CLAUDE_PROJECT_DIR/tools/meta_opt/log_event.sh\"",
+            "timeout": 5,
+            "async": true
+          }
+        ]
+      }
+    ],
+    "SessionEnd": [
+      {
+        "matcher": "",
+        "hooks": [
+          {
+            "type": "command",
+            "command": "bash \"$CLAUDE_PROJECT_DIR/tools/meta_opt/log_event.sh\"",
+            "timeout": 5,
+            "async": true
+          },
+          {
+            "type": "command",
+            "command": "bash \"$CLAUDE_PROJECT_DIR/tools/meta_opt/check_ready.sh\"",
+            "timeout": 5
+          }
+        ]
+      }
+    ]
+  }
+}

package/templates/gitignore-trace.txt

@@ -0,0 +1,3 @@
+# ARIS Review Tracing — add these lines to your project's .gitignore
+# Traces contain full LLM prompts/responses and should stay project-local.
+.aris/traces/