model-alignment-audit

model-alignment-audit is a skill for Claude Code, Codex from ngreenwall/claude-skills. It costs 147 tokens per session (1,480 once invoked), scanned A, original, MIT.

A review of skills and agent instruction files against Anthropic's current guidance and the behavior of newer models. It checks both file structure and whether the instructions still fit current models.

In plain words
What is it for?
Use it to audit SKILL.md, CLAUDE.md, AGENTS.md, and related context files for current-model fit.
Why use it?
Instructions can become outdated as models and recommended practices change, leading to unnecessary restrictions or poor results.

Skill for Claude CodeCodex

Written for no agent in particular: nothing here depends on one. Also seen: mentions CLAUDE.md; mentions subagents; mentions AGENTS.md.

Good fit Use it to audit SKILL.md, CLAUDE.md, AGENTS.md, and related context files for current-model fit.

Compare 6 skills from other repositories ↓
Install with agentmods
npx agentmods add skills/ngreenwall/claude-skills/model-alignment-audit
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 ngreenwall/claude-skills --skill model-alignment-audit
Clone the repo
git clone --depth 1 https://github.com/ngreenwall/claude-skills

Made for: Claude Code, Codex.

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 model-alignment-audit

README.md
[![agentmods](https://agentmods.dev/badge/skills/ngreenwall/claude-skills/model-alignment-audit/github.svg)](https://agentmods.dev/skills/ngreenwall/claude-skills/model-alignment-audit)
Your own site
<a href="https://agentmods.dev/skills/ngreenwall/claude-skills/model-alignment-audit"><img src="https://agentmods.dev/badge/skills/ngreenwall/claude-skills/model-alignment-audit/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 model-alignment-audit

Your own site · 80×15
<a href="https://agentmods.dev/skills/ngreenwall/claude-skills/model-alignment-audit"><img src="https://agentmods.dev/badge/skills/ngreenwall/claude-skills/model-alignment-audit.svg" alt="Reviewed on agentmods" width="80" height="20"></a>
Per session 147 Skills are progressive disclosure: only the name and description are preloaded; the body loads when the skill is used.
When invoked 1,480 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.
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 $0.00147 $0.01480
Opus 5 $0.00073 $0.00740
Sonnet 5 $0.00029 $0.00296
Haiku 4.5 $0.00015 $0.00148

Measured 8d ago against content hash 7a7f38829b3c, method: parsed. Prices are Anthropic first-party input rates as of 2026-09-08, from the pricing page.

Security

Grade A, and why

model-alignment-audit 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 8d 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.

model-alignment-audit/SKILL.md · 80 lines

How it starts

The opening of the file, as written. The whole thing — 80 lines — stays where its author put it; the contents beside it link to each section on GitHub.

SKILL: Model Alignment Audit

Checks whether skills and context files are still well-built for the models actually running them today, not just structurally valid. Two questions per file: does it follow Anthropic's current authoring guidance, and does its workflow logic (freedom levels, caps, hand-holding) still match what current models need, or is it carrying scaffolding tuned for older, less capable models.

This is NOT:

  • token-audit, that targets token-bloat cost (rewrite/split/merge). This skill targets guidance-and-model fit instead.
  • qa-audit, that finds contradictions and dead logic unrelated to model behavior. Overlap is possible, run both if both concerns apply.

Requires web access

This skill fetches live Anthropic documentation. If WebFetch/WebSearch are unavailable in this environment (e.g. blocked by org policy), tell the user and stop rather than proceeding without current guidance.

PROTOCOL

Step 1: Resolve targets

Default (no file named): audit everything in scope.

  • If run from a skills-library repo (a repo whose primary content is SKILL.md files): every SKILL.md in it, your global context file (if accessible from here), and this repo's CLAUDE.md.
  • If run from any other project: that project's CLAUDE.md (or AGENTS.md) plus any project-local skill files.

If the user names a specific file or skill, audit just that one instead. If the named file isn't a SKILL.md, CLAUDE.md/AGENTS.md, or a rules file, ask before proceeding.

If the file is managed by a sync tool that generates copies elsewhere (e.g. a tool that regenerates .claude/ or .cursor/ files from a single source of truth), always resolve to that source file, never a generated copy. Skip this check if you don't use such a tool, most projects edit SKILL.md/CLAUDE.md directly.

Step 2: Fetch current guidance

WebFetch https://platform.claude.com/docs/en/agents-and-tools/agent-skills/best-practices for the structural checklist. WebSearch for recent Sonnet/Opus release notes or changelogs, at most two fetches, summarize in under 200 words, to catch model-specific behavior changes relevant to freedom-level tuning, e.g. configurable reasoning effort, agentic capability jumps. Prefer anthropic.com/platform.claude.com sources over third-party summaries. If a fetch fails or returns nothing new, say so and fall back to references/checklist.md as-is rather than substituting search results. Summarize what's new since the checklist was last confirmed current, don't re-paste the full fetched pages into context.

Read the full file on GitHub · 80 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. 8d ago First seen · 80 lines · 147 tokens per session scan A 7a7f38829b3c

Subscribe to this mod's changes

model-alignment-audit is a skill published in the GitHub repository ngreenwall/claude-skills (2 stars, last pushed 9d ago), licensed MIT. It adds 147 tokens to every session and 1,480 once invoked, about $0.0007 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 2026-08-31.