Borrowing it
Nothing to install: this file belongs to kelesmert/turnlens. Take a copy, put it at the same path in your own repository, and replace the rules that are about this project with yours.
curl -O https://raw.githubusercontent.com/kelesmert/turnlens/main/.agents/skills/live-verification/SKILL.mdgit clone --depth 1 https://github.com/kelesmert/turnlensWrote 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/kelesmert/turnlens/live-verification)<a href="https://agentmods.dev/skills/kelesmert/turnlens/live-verification"><img src="https://agentmods.dev/badge/skills/kelesmert/turnlens/live-verification/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/kelesmert/turnlens/live-verification"><img src="https://agentmods.dev/badge/skills/kelesmert/turnlens/live-verification.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 | $0.00054 | $0.00588 |
| Opus 5 | $0.00027 | $0.00294 |
| Sonnet 5 | $0.00011 | $0.00118 |
| Haiku 4.5 | $0.00005 | $0.00059 |
Grade A, and why
live-verification 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 10d 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 — 58 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Live verification
A session file grows while it is being read. Any figure taken from a file that is still being written measures the growth, not the program. This has already produced one false result, and the rules below exist because of it.
Never verify against the session doing the work
Do not verify TurnLens against the agent session that is running the conversation. Point it at a separate, idle session instead.
On this machine that session is "Superpowers plugin installation"
(79c14101-4f5d-4793-ab69-8649e76d2062). Start or resume it in its own terminal,
drive it deliberately, and watch that one.
This binds verification, not use
The user running TurnLens on their own session, as a user, is the product working. Say nothing about it.
The rule binds only work done to check whether TurnLens is correct: comparing totals against an oracle, reproducing a bug, confirming a fix. If a watcher is running on the current session, that is worth raising only when a figure from it is about to be treated as evidence.
Copy a static file before measuring it
When a file has to be measured rather than watched, copy it first and measure the copy, so the bytes cannot change between two tools reading them.
Comparing across tools needs an isolated HOME as well. Pointing
CLAUDE_CONFIG_DIR at a frozen copy does not exclude the default roots:
ccusage scans both, deduplicates across them, and splits one session's usage
between two report entries. Under HOME=<frozen> the figures are unambiguous.
The false result this prevents
A live file was compared against a ccusage run taken minutes later, and the
resulting 0.33 percent gap was reported as a token-attribution difference. It was
the file growing during measurement.
Repeated on a frozen copy, the same rules reproduce ccusage exactly, with no
difference in any figure. Both runs are recorded in docs/VALIDATION.md.
Recording the result
A measurement is appended to docs/VALIDATION.md, never written over an earlier
one. A figure later found wrong is corrected by a new entry, because the original
figure is part of the record.
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.
- 10d ago First seen · 58 lines · 54 tokens per session scan A fe7ea4a91431
live-verification is a skill published in the GitHub repository kelesmert/turnlens (2 stars, last pushed 28d ago), licensed MIT. It adds 54 tokens to every session and 588 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 2026-08-31.
Other skills, from other repositories
grant-proposal-builder
Nonprofit and research grant proposal assembly with structured sections, budget templates, and evaluation alignment. Use when writing grant proposals, funding applications, or research funding requests.
accessibility-a11y
WCAG 2.2 compliance, ARIA patterns, keyboard navigation, screen readers, automated testing.
ai-ml-development
AI and machine learning development with PyTorch, TensorFlow, and LLM integration. Use when building ML models, training pipelines, fine-tuning LLMs, or implementing AI features.
android-development
Android development with Kotlin, Jetpack Compose, and modern Android architecture. Use when building Android apps, implementing Material Design, or following Android best practices.
audio-production
Professional audio production for music, podcasts, and sound design. Use when working with audio recording, mixing, mastering, or sound design for any medium.
authentication-patterns
OAuth 2.0, JWT, SSO, MFA, NextAuth/Clerk/Supabase Auth implementation patterns.