taste

taste is a skill for Claude Code, Codex from tomascupr/sous-chef. It costs 66 tokens per session (1,583 once invoked), scanned A, original, MIT.

A cross-model code-review process in which Codex reviews a code change and Claude checks each finding against the code.

In plain words
What is it for?
Use it for read-only reviews of working-tree changes, branches, or pull requests, especially when the change is large or risky.
Why use it?
It adds an independent review pass and removes findings that do not hold up when verified.

Skill for Claude CodeCodex

Written for Claude Code and Codex: shipped in a Claude Code plugin, but also runs codex exec. Also seen: mentions Codex.

Part of the sous-chef plugin — 7 skills shipped together

Good fit Use it for read-only reviews of working-tree changes, branches, or pull requests, especially when the change is large or risky.

Compare 6 skills from other repositories ↓
Install with agentmods
npx agentmods add skills/tomascupr/sous-chef/taste
Install

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.

Any agent
npx skills add tomascupr/sous-chef --skill taste
Clone the repo
git clone --depth 1 https://github.com/tomascupr/sous-chef

Made for: Claude Code, Codex.

Or install sous-chef, the plugin that ships this one along with the rest of its 7 skills.

Wrote 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.

agentmods badge for taste

README.md
[![agentmods](https://agentmods.dev/badge/skills/tomascupr/sous-chef/taste/github.svg)](https://agentmods.dev/skills/tomascupr/sous-chef/taste)
Your own site
<a href="https://agentmods.dev/skills/tomascupr/sous-chef/taste"><img src="https://agentmods.dev/badge/skills/tomascupr/sous-chef/taste/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.

agentmods 80×15 button for taste

Your own site · 80×15
<a href="https://agentmods.dev/skills/tomascupr/sous-chef/taste"><img src="https://agentmods.dev/badge/skills/tomascupr/sous-chef/taste.svg" alt="Reviewed on agentmods" width="80" height="20"></a>
Per session 66 Skills are progressive disclosure: only the name and description are preloaded; the body loads when the skill is used.
When invoked 1,583 The whole file, excluding the scripts and references it only reads on demand.
Security scan A 0 findings. A grade says what 26 rules found in the file — not that it is safe. Third-party audits
  • NVIDIA SkillSpector warn 7 Sept 2026
SkillSpector: 1 finding, up to medium

These are SkillSpector’s own severities. On a checked sample its high-severity flags on skills were ~96% false positives — a documented command, a public API, a “never do X” rule — so we show them as a caution to read, not a verdict. Why →

  • medium Rogue Agent · line 61
    Skill establishes unauthorized persistence across sessions via cron jobs, startup scripts, or state files. Session persistence allows an attacker to maintain access beyond the current interaction.
    Fix: Remove any persistence mechanisms (cron jobs, startup scripts, state files). Skills should not maintain state across sessions without explicit user consent.
How audits are shown
Origin original No closer match found in the catalogue.
Token cost

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.

ModelPer sessionOnce invoked
Fable 5.1 $0.00066 $0.01583
Opus 5 $0.00033 $0.00792
Sonnet 5 $0.00013 $0.00317
Haiku 4.5 $0.00007 $0.00158

Measured 9d ago against content hash 39cbc083a59c, method: parsed. Prices are Anthropic first-party input rates as of 2026-09-08, from the pricing page.

Security

Grade A, and why

taste 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 9d 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.

skills/taste/SKILL.md · 72 lines

How it starts

The opening of the file, as written. The whole thing — 72 lines — stays where its author put it; the contents beside it link to each section on GitHub.

Taste - the chef tastes every plate before it leaves the kitchen

A reviewer from a different training lineage doesn't share your blind spots - cross-model review reliably catches edge-case bugs that self-review misses. But raw Codex findings over-flag, so nothing reaches the user unvalidated: the deliverable of a taste is the CONFIRMED set, written to disk where a refire can pick it up.

If codex is missing or ~/.codex/sous-chef.config.toml doesn't exist, stop and offer /sous-chef:mise (Codex silently ignores a missing profile - check with test -f, don't rely on an error).

1. Scope the diff

git status --short --untracked-files=all
git diff --shortstat && git diff --shortstat --cached
  • Working-tree review (default): staged + unstaged + untracked count as reviewable.
  • Branch review (user said "review this branch/PR"): scope with git diff <base>...HEAD.
  • Post-fire review (a pre-fire baseline exists - inside /sous-chef:serve, or tasting a fire's result on a tree that was dirty before the fire): the change under review is the delta since that baseline, never the whole tree. Scope the prompt to it - name the files the run touched, and point at the baseline patch for files that mix fired changes with prior WIP. Findings against pre-existing WIP would send a refire rewriting the user's own uncommitted work; say in the report that WIP was excluded.
  • Empty scope → say there's nothing to review; don't run Codex on nothing.
  • Large diffs review shallow. If the scope exceeds ~1,500 changed lines, split into per-area passes or agree a focus with the user - and say that's why.

2. Run the review - read-only, in the background

Mint a job dir (JOB=$(mktemp -d "$SCRATCHPAD/taste-XXXXXX"), where $SCRATCHPAD is your session scratchpad directory - substitute its absolute path). Build the review prompt from references/review-prompt.md, filling in the diff scope and any focus the user gave, then:

Bash (run_in_background: true), cwd = repo root:
env -u CODEX_API_KEY -u CODEX_ACCESS_TOKEN codex exec --profile sous-chef --sandbox read-only \
  --output-last-message "$JOB/result.md" \
  - < "$JOB/review-prompt.md" > "$JOB/job.log" 2>&1

Read the full file on GitHub · 72 lines

Files

What ships with it

1 file beside SKILL.md in the same directory: the scripts, references and assets a skill reads on demand. Not counted in the per-session cost; read them before you install if any of them is executable.

Changes

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.

  1. 9d ago First seen · 72 lines · 66 tokens per session scan A 39cbc083a59c

Subscribe to this mod's changes

taste is a skill published in the GitHub repository tomascupr/sous-chef (76 stars, last pushed 1mo ago), licensed MIT. It adds 66 tokens to every session and 1,583 once invoked, about $0.0003 per session on Opus 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 2026-08-30.

Related

Other skills, from other repositories