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.
npx skills add NickCrew/Claude-Cortex --skill ai-tells-reviewgit clone --depth 1 https://github.com/NickCrew/Claude-CortexWrote 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/skills/nickcrew/claude-cortex/ai-tells-review)<a href="https://agentmods.dev/skills/nickcrew/claude-cortex/ai-tells-review"><img src="https://agentmods.dev/badge/skills/nickcrew/claude-cortex/ai-tells-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/skills/nickcrew/claude-cortex/ai-tells-review"><img src="https://agentmods.dev/badge/skills/nickcrew/claude-cortex/ai-tells-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 | $0.00075 | $0.02242 |
| Opus 5 | $0.00037 | $0.01121 |
| Sonnet 5 | $0.00015 | $0.00448 |
| Haiku 4.5 | $0.00007 | $0.00224 |
Grade A, and why
ai-tells-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 7d 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.
How it starts
The opening of the file, as written. The whole thing — 159 lines — stays where its author put it; the contents beside it link to each section on GitHub.
AI Tells — Structural Review
A judgment-based pass that catches the subtle tells the surface scan misses. Cite each finding, name the family, suggest a fix-direction. Then offer to rewrite.
When To Use
- Public-facing artifacts: READMEs, shipped docs, blog posts, release notes
- Technical writing where the prose itself is the deliverable
- After
ai-tells-scanhas already stripped the obvious tokens - When a draft "reads AI" but you can't articulate why
- Before publishing anything Nick wouldn't want attributed to a machine
Skip when: the artifact is internal-only and nobody cares, or the user only wants the mechanical pass (ai-tells-scan is enough).
Workflow
- Read the artifact end to end. Don't pattern-match line by line yet. Get the rhythm and shape first.
- Run the four-family rubric below. For each tell found, note: line(s), family, why it reads as a tell, fix-direction.
- Produce the findings report in the structure under "Output Shape."
- Offer the rewrite explicitly: "Want me to do a pass that fixes these?"
- If yes, rewrite preserving structure and content. Don't change what the doc is saying — change how it's saying it.
The user can stop at the report. Don't rewrite without explicit consent.
The Four-Family Rubric
1. Structural Reflexes
Default shapes the model reaches for regardless of fit.
- Aphoristic closer. Sentence trying to land like a punchline at the end of a section. Examples: "Premature alerts produce numbness." / "the auditor is always ahead." / "No artefact, no promotion." Fix-direction: end on the substantive point, not the epigram.
- Definitional opener. Every section starts "X is a Y that…" even when X was already established. Fix-direction: start with the action or the question instead.
- Echo paragraph. Paragraph N+1 restates paragraph N with synonyms. Fix-direction: delete one. Usually the second.
- Summary-of-summary closer. "In essence…" / "Ultimately…" wrapping up what was just said. Fix-direction: delete the closer; the section already ended.
- Over-signposting. "There are three things to consider. First…" when headers already do this. Fix-direction: delete the meta-narration.
- Self-narration. Prose commenting on what the doc is doing: "That's the dashboard's reason for existing." / "(this is where we start)" inside a diagram. Fix-direction: let the doc do the thing instead of describing itself.
- Decision-stamp formality.
**Decided 2026-05-18:**court-ruling style in personal notes. Fix-direction: drop the stamp; state the decision in prose. - Labeled callback scheme. Q1–Q6, Layer 1/2/3, deliberately numbered cross-refs in a doc that didn't need them. Fix-direction: use names, not numbers; let prose carry the references.
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.
- 7d ago First seen · 159 lines · 75 tokens per session scan A 17eb9ccd2ef7
ai-tells-review is a skill published in the GitHub repository NickCrew/Claude-Cortex (40 stars, last pushed 2mo ago), licensed MIT. It adds 75 tokens to every session and 2,242 once invoked, about $0.0004 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 2026-09-03.
Other skills, from other repositories
template-prose-project
Prose-review pipeline exemplar — readability gates, structural checking, BibTeX validation, and quality review workflows.
etincel-nonfiction
Write or revise non-fiction prose (emails, essays, blog posts, memos, newsletters, docs, long-form articles) so it doesn't read as AI-generated, in the user's own trained voice or a chosen preset tone. Use this any time the user is drafting or revising non-fiction prose of meaningful length inside Claude Code, Claude…
humanizer
Humanize text: strip AI-isms and add real voice.
Agent Instruction Drift Check
Procedure for comparing an agent-facing instruction doc (AGENTS.md/CLAUDE.md-style) against the actual current code and configuration to find and flag mismatches.
oiticica-clarity-faults
Audit a single English passage or compare parallel drafts using the seven component skills in Oiticica's faults-of-clarity section.
oiticica-style
Orchestrate Oiticica aggregate audits for English prose by identifying genre, eliminating defects, then applying qualities.