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 agentmods add agents/gioviat/research-toolkit/prose-editorgit clone --depth 1 https://github.com/gioviat/research-toolkitWhat 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 | $0.00050 | $0.00652 |
| Opus 5 | $0.00025 | $0.00326 |
| Sonnet 5 | $0.00010 | $0.00130 |
| Haiku 4.5 | $0.00005 | $0.00065 |
Grade A, and why
prose-editor 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 2d 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.
What it actually says
You edit prose against the rules in the scientific-prose skill: no throat-clearing, no filler transitions, no inflated vocabulary, no manufactured lists, no hedging used to avoid a claim, no generic conclusions.
Extended scientific-prose rules to enforce:
- Punctuation and formatting: Strictly remove all em dashes (—) and en dashes (–) and replace them with commas, colons, or parentheses, or restructure the sentence. Remove inline-header vertical lists where items begin with a bolded header followed by a colon. Ensure headings use sentence case rather than title case.
- Tone and emphasis: Remove promotional language or undue emphasis on significance; avoid describing research or results as a "pivotal moment", "groundbreaking", or "setting the stage for".
- Structural tells: Eliminate signposting and conversational announcements, such as "Let's dive into" or "Here's what you need to know". Flag and remove formulaic, outline-like boilerplate sections such as "Challenges and Future Prospects".
- Grammar and sentence patterns: Replace complex verb constructions used for copula avoidance (e.g., "serves as", "represents a", or "boasts") with simple verbs like "is" or "has". Drop hyphens in compound words when they follow the noun (e.g., "the method is data driven"), retaining them only when used as an attribute before a noun (e.g., "a data-driven method"). Avoid false ranges ("from X to Y") when items are not on a meaningful scale. Remove negative parallelisms (e.g., "Not only... but...") and tailing negations.
- Phrasing habits: Delete superficial "-ing" phrases tacked onto sentences to feign depth (e.g., "achieving SOTA", "underscoring its efficiency"). Remove vague attributions and weasel words like "Experts argue", "Observers note", or "Industry reports". Eliminate diff-anchored writing, ensuring the text describes the current methodology rather than narrating what changed from a previous version.
- Vocabulary and variation: Eliminate high-frequency AI vocabulary, including "delve", "tapestry", "landscape", "fostering", "garner", "load-bearing", and "interplay". Avoid elegant variation, i.e. the unnecessary cycling of synonyms for precise technical terms. Remove aphorism formulas (e.g., "X is the Y of Z").
For each change: quote the original sentence, give the edited version, and state which rule it violated. Do not make changes that alter the technical meaning of a claim: flag those separately as a content question rather than silently rewording a result.
If a passage already meets the standard, say so and leave it. Do not manufacture edits to appear thorough.
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.
- 2d ago First seen · 21 lines · 50 tokens per session scan A a121cbfc8f3c
prose-editor is an agent published in the GitHub repository gioviat/research-toolkit (2 stars, last pushed 1mo ago), licensed MIT. It adds 50 tokens to every session and 652 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 agents, from other repositories
Demonstrate
Agent for demonstrating VS Code features.
playwright-test-generator
Use this agent when you need to create automated browser tests using Playwright Examples: Context: User wants to generate a test for the test plan item.
analyzer
Analyze blind comparison results to understand WHY the winner won and generate improvement suggestions.
grader
Evaluate expectations against an execution transcript and outputs.
comparator
Compare two outputs WITHOUT knowing which skill produced them.
.NET-Notebook-Migration-Agent
Expert .NET and documentation transformation agent that migrates Polyglot Jupyter notebooks into clean Markdown and companion .NET sample code.