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/w2ur/claude-code-setupWrote 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/w2ur/claude-code-setup/model-review)<a href="https://agentmods.dev/commands/w2ur/claude-code-setup/model-review"><img src="https://agentmods.dev/badge/commands/w2ur/claude-code-setup/model-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/w2ur/claude-code-setup/model-review"><img src="https://agentmods.dev/badge/commands/w2ur/claude-code-setup/model-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 | $0.00047 | $0.01514 |
| Opus 5.5 | $0.00019 | $0.00606 |
| Sonnet 5.5 | $0.00009 | $0.00303 |
| Haiku 4.5 | $0.00005 | $0.00151 |
Grade B, and why
model-review scanned grade B with 1 finding 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 6d 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.
Reads agent configuration directoriesmediumAgent snooping
.claude/, .codex/, .gemini/ hold keys, settings and other credentials a mod has no legitimate need for.
jq '.modelSettings' ~/.claude/settings.json How it starts
The opening of the file, as written. The whole thing — 111 lines — stays where its author put it; the contents beside it link to each section on GitHub.
A new model has shipped and the owner wants the setup checked against it. The checklist below is fixed; your judgment goes into what each hit means, not into which files to look at.
$0 is the new model id. $1, if given, is the model it replaces. If $0 is empty, ask
for it — do not guess from the session model.
Step 0 — the standing decisions are not re-opened
Read the Sub-agents section of ~/.claude/CLAUDE.md: the tier lattice, which tier does
what, and that effort lives per model in settings.json modelSettings. Apply it; do not
copy it into your report and do not propose changing it. A new release changes which model
an alias resolves to, not the lattice.
Step 1 — derive the targets. Never hand-list them
Run this block exactly, as one single Bash call: shell state does not survive between calls, so a grep split into its own call runs with no targets. It works under zsh and bash. The target set is whatever it prints today:
targets=("$HOME/.claude/scripts")
while IFS= read -r f; do targets+=("$f"); done \
< <(find "$HOME/Library/LaunchAgents" -maxdepth 1 -name 'com.example.*.plist')
for d in "$HOME"/Dev/*/; do
case "$(basename "$d")" in my-encyclopedia-app|my-bible-app|my-facilitation-app) continue ;; esac
for s in scripts src options docs/triggers .claude/agents; do
[ -e "$d$s" ] && targets+=("$d$s")
done
done
ex=(--exclude-dir=node_modules --exclude-dir=.venv --exclude-dir=.git
--exclude-dir=.next --exclude-dir=dist)
printf '%s\n' '--- 1. model pins'
grep -rnE 'claude-(opus|sonnet|haiku|fable)-[0-9]|--model|^model: *(opus|sonnet|haiku|fable)' \
"${ex[@]}" "${targets[@]}"
printf '%s\n' '--- 2. failing control'
real=$(grep -rhE 'claude-(opus|sonnet|haiku|fable)-[0-9]' "${ex[@]}" "${targets[@]}" | wc -l)
typo=$(grep -rhE 'claude-(opsu|sonet|haiky|fabel)-[0-9]' "${ex[@]}" "${targets[@]}" | wc -l)
if [ "$real" -gt 0 ] && [ "$typo" -eq 0 ]; then
printf 'control ok: real=%d misspelled=%d\n' $real $typo
else
printf 'CONTROL FAILED: real=%d misspelled=%d — the probe proves nothing\n' $real $typo
fi
printf '%s\n' '--- 3. headless claude -p calls'
grep -rnE '(claude|CLAUDE_BIN\}?"?) +-p( |$)' "${ex[@]}" "${targets[@]}"
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.
- 6d ago Changed fbae28b32544
- 21d ago First seen · 111 lines · 47 tokens per session scan B 7ea5e0a1b8b0
model-review is a command published in the GitHub repository w2ur/claude-code-setup (2 stars, last pushed 3d ago), licensed MIT. It adds 47 tokens to every session and 1,514 once invoked, about $0.0002 per session on Opus 5.5. A static security scan graded it B with 1 finding (reads agent configuration directories). No closer match exists in the catalogue, so it is treated as the original; first seen 2026-09-26.
Other commands, from other repositories
loop
Autonomous execution loop. Runs the TDD cycle repeatedly until all tasks in an epic are complete, with circuit breaker protection and PR workflow.
execute
Execute the next available tracked task with TDD, pre-commit validation, PR creation, AI code review, and issue tracker bridge sync.
audit-agents-skills
Audit quality of agents, skills, and commands in a Claude Code project.
land-and-deploy
Merge PR, wait for CI, verify deploy, run canary — the complete landing pipeline.
dashboard
Compact visual progress dashboard with progress bar, task table, dependency graph, and active/blocked/next status.
check-cache-bugs
Audit Claude Code setup for cache bugs (CC#40524) — sentinel, --resume/--continue, attribution header + ArkNill B3/B4/B5.