Getting it into your agent
One page per mod, every tool's command on it. A separate URL per tool would split the same page into five that compete with each other.
git clone --depth 1 https://github.com/topprismdata/cultivating-ml-agentWrote this? Show the measurements
A badge with what this costs and how it scanned, read live from this page, so it follows the numbers instead of freezing them. Markdown for a README, HTML for a documentation site or a project page.
[](https://agentmods.dev/commands/topprismdata/cultivating-ml-agent/review)<a href="https://agentmods.dev/commands/topprismdata/cultivating-ml-agent/review"><img src="https://agentmods.dev/badge/commands/topprismdata/cultivating-ml-agent/review/github.svg" alt="Measured on agentmods" height="20"></a>Or the 80×15 button, for a site that already has a row of RSS and ATOM ones. Only the verdict fits; the numbers stay here.
<a href="https://agentmods.dev/commands/topprismdata/cultivating-ml-agent/review"><img src="https://agentmods.dev/badge/commands/topprismdata/cultivating-ml-agent/review.svg" alt="Reviewed on agentmods" width="80" height="20"></a>What it costs to keep this loaded
Counted locally with the o200k_base tokenizer, which is exact for GPT models; Claude uses its own tokenizer and its counts differ. Treat this as one consistent yardstick across the catalogue rather than a bill. Prices are per million input tokens.
| Model | Per session | Once invoked |
|---|---|---|
| Fable 5.1 | $0.00025 | $0.00434 |
| Opus 5.5 | $0.00010 | $0.00174 |
| Sonnet 5.5 | $0.00005 | $0.00087 |
| Haiku 4.5 | $0.00003 | $0.00043 |
Grade A, and why
review scanned grade A with 0 findings against 26 rules in 11 categories — prompt injection, anti-refusal, data exfiltration, privilege escalation, supply chain, agent snooping, system-prompt leakage, SSRF and excessive agency — measured 2mo ago.
A static scan of the body, not an audit. Every finding is printed with the line that produced it so you can judge whether it matters here. A mod is markdown that instructs an agent; that is exactly why what it instructs is worth reading.
Nothing flagged
None of the 26 patterns this scan looks for appear in this file: no shell pipes, no recursive deletes, no credential paths, no hidden text, no instruction-override or anti-refusal phrasing, no agent-config snooping. That is not a guarantee, it is the absence of the things that are checkable.
What it actually says
/review — Cross-Model Review
Use an external LLM as a critic. Breaks self-play blind spots. Auto-detects available backend.
Usage
/review <topic>
/review <topic> --backend=agy
Examples:
/review should I add CatBoost to the TPS May 2022 stack?/review is this the right CV strategy for time-series data?/review before I submit my final submission for jigsaw-toxic
What This Does
- Summarizes your current approach + key claims
- Invokes
cross_review.sh(auto-detects: agy > gemini > codex > ollama) - Captures the critique
- Synthesizes: what to keep, what to reject, what's new
- Saves the review to
memory/cross-reviews/<topic-slug>.md - Adds an entry to MEMORY.md
Backend Auto-Detection
The script checks in this order:
agy(Antigravity CLI) — preferredgemini(Google Gemini CLI)codex(OpenAI Codex CLI)ollama(local models, no API)
If none installed, falls back to adversarial self-check (5-question rubric).
When to Use This
- Before submitting a final submission
- Before making an architectural decision
- After 3+ failed attempts at the same problem
- When you suspect your reasoning is going in circles
When NOT to Use
- Trivial decisions (not worth the latency)
- Decisions where you've already gotten external input
- When the answer is clearly defined (e.g., "use GroupKFold for groups")
Anti-Patterns
- ❌ Using /review to validate (asking for confirmation, not critique)
- ❌ Ignoring the critique
- ❌ Not documenting the outcome
What this file has done since we first saw it
Hashed on every crawl. A supply-chain change to an agent config is a question of when, not whether, so the history is kept rather than the latest state alone.
- 2mo ago First seen · 58 lines · 25 tokens per session scan A d01cf602da73
review is a command published in the GitHub repository topprismdata/cultivating-ml-agent (5 stars, last pushed 25d ago), licensed MIT. It adds 25 tokens to every session and 434 once invoked, about $0.0001 per session on Opus 5.5. A static security scan graded it A with 0 findings. No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-31.
review compared with commands from other repositories
| Mod | Type | Stars | Last push | Licence | Scan | Tokens | Works with |
|---|---|---|---|---|---|---|---|
| reviewtopprismdata/cultivating-ml-agent | Command | 5 | 25d ago | MIT | A | 25 | Claude Code |
| adversarial-reviewAgentToolkit/altk-evolve | Command | 135 | 3d ago | Apache-2.0 | A | 26 | Claude Code |
| gsd-shipSpillwaveSolutions/agent-brain | Command | 119 | 1mo ago | MIT | A | 13 | Claude Code, OpenCode |
| slm-graphqualixar/superlocalmemory | Command | 273 | 9d ago | AGPL-3.0 | A | 4 | GitHub Copilot |
| unforgit-curateMiguelMedeiros/unforgit | Command | 11 | yesterday | MIT | A | 0 | Claude Code |
| speckit.analyzeSpillwaveSolutions/agent-brain | Command | 119 | 1mo ago | MIT | A | 26 | Claude Code |
| scankage-core/Kage | Command | 41 | 2mo ago | GPL-3.0 | A | 27 | Claude Code |