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 kalyvask/winning-writing --skill fact-checkergit clone --depth 1 https://github.com/kalyvask/winning-writingWrote 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/kalyvask/winning-writing/fact-checker)<a href="https://agentmods.dev/skills/kalyvask/winning-writing/fact-checker"><img src="https://agentmods.dev/badge/skills/kalyvask/winning-writing/fact-checker/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/kalyvask/winning-writing/fact-checker"><img src="https://agentmods.dev/badge/skills/kalyvask/winning-writing/fact-checker.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 | $0.00000 | $0.01443 |
| Opus 5 | $0.00000 | $0.00722 |
| Sonnet 5 | $0.00000 | $0.00289 |
| Haiku 4.5 | $0.00000 | $0.00144 |
Grade A, and why
fact-checker 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 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.
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 — 90 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Fact checker
Source: points/ai-writing-rules.md (section 2, Accuracy), points/exec-memo-rules.md (rules 4 and 7, on numbers and citation), points/named-failure-modes.md.
Why this exists
Every other skill in this repo makes the writing better. This one keeps it true. Those are different jobs and the second one is worth more, because style problems cost you a reader and accuracy problems cost you the relationship.
One invented number, one misattributed quote, one wrong job title — and every other number in the document becomes suspect. Credibility is a budget you spend once.
The two failure modes this catches:
- The model made it up. LLMs fabricate confidently: invented statistics, plausible-sounding quotes attributed to real people, dates that fit the narrative. Fluency is not evidence.
- You knew it once and it changed. Headcounts, funding rounds, job titles, pricing, product names. True in March, wrong in August. Stale facts read exactly like fabricated ones to the person who knows.
What counts as a checkable claim
Extract every instance of:
| Type | Examples |
|---|---|
| Numbers | percentages, dollar amounts, headcounts, dates, durations, rankings, sample sizes |
| Named people | spelling, current employer, current title, pronouns if stated |
| Organizations | company name spelling, what they actually do, funding, ownership |
| Quotes | exact wording, who said it, where, when |
| Citations | does the linked source exist, and does it say what you claim |
| Superlatives | "first," "only," "largest," "fastest" — these are factual claims, not adjectives |
| Implied claims | "since we launched" (did you?), "our fastest-growing segment" (is it?) |
Superlatives and implied claims are the ones writers forget are checkable. Treat them as claims.
Procedure
- Extract. List every checkable claim as a separate row before verifying anything. Quote the exact span from the draft. Do not paraphrase — the wording is what ships.
- Classify the source of each claim. One of: user-supplied, model-generated, or inherited from an earlier draft. Model-generated claims get the harshest scrutiny — that is where fabrication lives.
- Verify. Use web search for anything public. For internal claims (your own metrics), you cannot verify — say so, and ask the user for the source.
- Report in the table below.
- Rewrite the failures. For every ✗ and every ?, propose the specific replacement line.
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 · 90 lines · 0 tokens per session scan A ae135681c953
fact-checker is a skill published in the GitHub repository kalyvask/winning-writing (14 stars, last pushed 6d ago), licensed MIT. It costs nothing until one of its globs matches a file; then it loads 1,443 tokens. 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 2026-08-30.
Other skills, from other repositories
project-memory
Generate a project-specific context file from a brief so an AI assistant remembers your editorial constraints, voice, audience, and quality bar across sessions.
project-retrospective
Generate a LESSONS.md from a finished project: what worked, what didn't, what to reuse, what to retire — formatted for next-project carry-over.
template-selector
Recommend the right skill bundle, agent, and workflow sequence for a new project — so media professionals can start producing instead of browsing a 394-skill library.
multi-author-harmonizer
Reviews a text written or assembled by multiple authors and produces a detailed inconsistency report — flagging voice shifts, terminology mismatches, tonal clashes, and formatting discrepancies — with specific harmonisation recommendations for each.
jargon-flagger
Scans a draft and flags every instance of technical jargon, unexplained acronyms, and insider language that a general-audience reader would not understand — with a plain-language alternative for each.
passive-voice-checker
Identifies every passive-voice construction in a draft, assesses whether each weakens or serves the prose, and offers active-voice alternatives for those that should be changed — while leaving justified passives alone.