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 niels-emmer/myace --skill data-classification-guidegit clone --depth 1 https://github.com/niels-emmer/myaceWrote 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/niels-emmer/myace/data-classification-guide)<a href="https://agentmods.dev/skills/niels-emmer/myace/data-classification-guide"><img src="https://agentmods.dev/badge/skills/niels-emmer/myace/data-classification-guide/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/niels-emmer/myace/data-classification-guide"><img src="https://agentmods.dev/badge/skills/niels-emmer/myace/data-classification-guide.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 | $0.00032 | $0.00897 |
| Opus 5.5 | $0.00013 | $0.00359 |
| Sonnet 5.5 | $0.00006 | $0.00179 |
| Haiku 4.5 | $0.00003 | $0.00090 |
Grade A, and why
Data Classification Guide 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 1mo 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 — 52 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Purpose
Give reviewers a shared, concrete way to decide how carefully a piece of data needs to be handled, instead of relying on a case-by-case judgment call every time. Most security and privacy findings ("this shouldn't be in the log," "this shouldn't be in the prompt") come down to misjudging which tier the data actually belongs to.
When to use it
Whenever a review needs to answer "is it OK for this data to end up here" — in an AI prompt or context window, in an application or access log, in a committed file (including test fixtures and example config), or in an error message returned to a client. Use it alongside security-audit-checklist's injection/output-handling items and the Data Classification Awareness rule.
The three tiers
Public — safe for anyone to see with no restriction: published documentation, open-source code, marketing content, anything already deliberately made public.
Internal — not secret, but not meant for outside distribution: internal architecture notes, non-sensitive business metrics, internal tooling config that doesn't grant access to anything, employee directory info the org treats as internal-only.
Sensitive — data whose exposure causes real harm: personally identifiable information (names tied to contact info, government IDs, dates of birth), authentication material (passwords, API keys, tokens, private keys, session identifiers), financial data (card numbers, account numbers, transaction details), health data, and anything a contract, regulation, or the project's own policy specifically restricts.
When a piece of data's tier isn't obvious, classify it at the higher (more restrictive) tier until someone with authority over the data confirms otherwise — the cost of over-protecting public-adjacent data is small; the cost of under-protecting sensitive data is not.
What's allowed where, by tier
| Destination | Public | Internal | Sensitive |
|---|---|---|---|
| AI prompt / context window | Yes | Generally yes, if the tool/provider is already trusted with internal data | No, unless the specific tool is explicitly approved for that data class and the org has confirmed it — default to no |
| Application/access logs | Yes | Yes, if log access is itself restricted to internal staff | No — never log raw sensitive values; log a reference/ID instead if you need traceability |
| Committed files (repo, fixtures, examples) | Yes | Usually no — internal specifics don't belong in a public or widely-shared repo | Never — including "just for a test," "temporarily," or in a private repo (private repos still get cloned, forked, and mirrored) |
| Client-facing error messages | Yes | No | No |
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.
- 1mo ago First seen · 52 lines · 32 tokens per session scan A a1e90dd7aa88
Data Classification Guide is a skill published in the GitHub repository niels-emmer/myace (1 stars, last pushed 3d ago), licensed MIT. It adds 32 tokens to every session and 897 once invoked, about $0.0001 per session on Opus 5.5. 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
idiomatic-go
Apply Go community idioms for API and package design, error and concurrency patterns, naming (initialisms, stutter, receivers), generics vs interfaces, and Go code review — including PR or diff review, questions about "the Go way", pointer vs value receivers, channels vs mutexes, and building Go language tooling…
eclipse-license-check
Vet third-party dependencies for an Eclipse Foundation project with the Eclipse Dash License Tool — license checks, "do I need to vet X?" questions, IPLab review filing, and release IP due diligence. Use whenever the user mentions Eclipse license checking, dash-licenses, the Dash License Tool, IPLab, third-party…
go-documentation
Write idiomatic Go doc comments that render correctly on pkg.go.dev — package comments, doc.go, exported-symbol comments, testable Example functions, deprecation notices — and publish or debug modules on pkg.go.dev (pkgsite, the module proxy). Use when writing or reviewing Go documentation, or when a developer used to…
publish-vscode-extension
Publish a VS Code extension to the VS Code Marketplace and/or Open VSX Registry. Use when the user wants to publish, ship, or release a VS Code extension; audit or improve manifest, listing, or release setup; build or upload a .vsix; set up vsce or ovsx; create a marketplace publisher or Open VSX namespace; configure…
domainmodel-dsl
Read, write, and modify Domain Model DSL files (.dmodel) — datatypes, entities with inheritance and multi-valued features, and nested packages with qualified cross-references. Use when the user wants to create, edit, explain, or debug a domain model in this DSL, including resolving references across packages.
statemachine-dsl
Read, write, and modify Statemachine DSL files (.statemachine). Use when the user wants to create, edit, explain, or debug a state machine written in this textual DSL — defining states, events, commands, and transitions, or generating a machine from a description.