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 Dhirenderchoudhary/unmarkk --skill unmark-textgit clone --depth 1 https://github.com/Dhirenderchoudhary/unmarkkWrote 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/dhirenderchoudhary/unmarkk/unmark-text)<a href="https://agentmods.dev/skills/dhirenderchoudhary/unmarkk/unmark-text"><img src="https://agentmods.dev/badge/skills/dhirenderchoudhary/unmarkk/unmark-text/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/dhirenderchoudhary/unmarkk/unmark-text"><img src="https://agentmods.dev/badge/skills/dhirenderchoudhary/unmarkk/unmark-text.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 | $0.00083 | $0.01036 |
| Opus 5 | $0.00042 | $0.00518 |
| Sonnet 5 | $0.00017 | $0.00207 |
| Haiku 4.5 | $0.00008 | $0.00104 |
Grade C, and why
unmark-text scanned grade C with 1 finding 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 12d 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.
Hidden instructionshighPrompt injection
Directives inside HTML comments, invisible characters or bidirectional overrides are read by the model and not by the person reviewing the file.
- **Emoji and script joiners are preserved.** 👨👩👧 is six codepoints, two of them invisible, and Persian میروم needs its zero-width non-joiner. The tool keeps invisibles that sit after a base they belong to and remo How it starts
The opening of the file, as written. The whole thing — 115 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Text hygiene
Two different problems, in order of how reliably they can be solved.
Invisible characters get into text by accident far more often than by design: a bad copy-paste, a CMS that inserts soft hyphens, a word processor using non-breaking spaces for layout. They break search, diffs, screen readers and string comparison. Removing them is deterministic and safe.
Statistical watermarks are the harder case, and the honest answer is covered at the bottom.
The deterministic pass
unmark inspect draft.md # what is in there
unmark clean draft.md -o clean.md # remove it
Or in a pipeline:
cat draft.md | unmark clean - > clean.md
Inspect before cleaning when editing an existing file. Show the user what was found; it is usually more interesting than the removal.
Defaults worth knowing
- Exotic spaces are normalised to a plain space. For prose that is right.
For anything where a non-breaking space is doing layout work — a table, a
measurement like
5 kg, a French guillemet — pass--keep-spaces. - Emoji and script joiners are preserved. 👨👩👧 is six codepoints, two of
them invisible, and Persian میروم needs its zero-width non-joiner. The tool
keeps invisibles that sit after a base they belong to and removes the ones
floating free. Only use
--strip-emoji-glueif the user explicitly wants every invisible character gone and accepts that it breaks those. --aggressiveand--nfkcare off for a reason. The first rewrites Cyrillic and fullwidth lookalikes to ASCII, which mangles multilingual text. The second changes visible characters. Ask before either.
Protect the non-prose
When prose and code are mixed, the deterministic pass is safe on the whole file — it never touches ASCII. But if you are rewriting rather than cleaning, leave these alone byte for byte:
- fenced and inline code
- commands, paths, URLs, identifiers, API names
- formulas, citations, and anything quoted verbatim
- required legal, academic or platform disclosures
What ships with it
2 files 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.
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.
- 12d ago First seen · 115 lines · 83 tokens per session scan C ff03fce5f78a
unmark-text is a skill published in the GitHub repository Dhirenderchoudhary/unmarkk (6 stars, last pushed 12d ago), licensed MIT. It adds 83 tokens to every session and 1,036 once invoked, about $0.0004 per session on Opus 5. A static security scan graded it C with 1 finding (hidden instructions). No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-31.
Other skills, from other repositories
create-pr
Creates a GitHub PR with a Linear-ticket-prefixed title and a decision-led, narrative description for Prisma 8. Use when the user wants to create a pull request, open a PR, or submit changes for review.
review-implement-phase
Implements triaged review actions, commits focused fixes, and posts Done plus resolves threads. Use when the user wants only the implementation phase of the review-framework workflow.
no-bare-casts
Writing as in TypeScript or TSX production code, modifying a file that contains a bare as cast, silencing a type error with a cast, encountering as unknown as, or reviewing a cast site.
write-architecture-docs
Write or rewrite architecture subsystem docs, ADRs, and reference material for the engineering team. Use when creating, updating, or reviewing docs under docs/architecture docs/, or when the user asks you to write documentation that describes the system's design.
pr-checklist
Use when opening or finalizing a GitHub PR for OpenMetadata. Walks through the repo PR template — linked issue, high-level design (for big PRs), unit/integration/Playwright tests + coverage, UI screen recording, and manual test steps — then drafts a fully-filled PR body and (optionally) creates the PR.
systematic-debugging
Use when debugging a failing test, build error, or runtime issue that isn't immediately obvious. Guides a 4-phase root cause analysis instead of random fix attempts.