metis-strategy/metis-whitepaper
Produce Metis Strategy whitepapers — 15–28 page thought leadership PDFs with a specific structure, visual style, and editorial voice. Use this skill whenever a user asks to create, draft, or structure a whitepaper, thought leadership paper, playbook, or long-form research report for Metis Strategy. Also trigger when the user mentions "whitepaper," "white paper," "playbook," "thought leadership piece," "long-form report," or asks to produce a publishable document that looks like existing Metis research. Output is a polished, print-ready PDF generated from HTML via Playwright — matching the look and feel of published Metis whitepapers. Always use this skill for whitepaper requests, even if the user just says "write a whitepaper on X" without further specification.
74
93%
Does it follow best practices?
Run evals on this skill
Adds up to 20 points to the overall score
View guide
Low
Low-risk findings worth noting
Low
Low-risk findings.
2 low severity findings. Worth noting, but not necessarily harmful.
W011: Third-party content exposure detected (indirect prompt injection risk).
What this means
The skill exposes the agent to untrusted, user-generated content from public third-party sources, creating a risk of indirect prompt injection. This includes browsing arbitrary URLs, reading social media posts or forum comments, and analyzing content from unknown websites.
Why it was flagged
The SKILL.md explicitly requires conducting external web research and fetching/verifying public URLs (Step 1: "Perform web research...collect sources with URLs" and Step 8: "fetch each endnote URL...pull a verbatim quote"), including open/public sources like Stack Overflow, GitHub, arXiv, industry analysts and government sites, so the agent will read and act on untrusted third‑party content that can materially influence drafting and tool actions.
Where we found it
survey.stackoverflow.co
domain · 3 sites
SKILL.md Step 8 Part B explicitly instructs the agent to fetch each endnote URL and pull verbatim quotes; survey.stackoverflow.co/2025/ is the concrete example URL used in the endnote template, meaning the agent will read and act on content from this third-party domain.
SKILL.md
492
https://survey.stackoverflow.co/2025/
references/content-patterns.md
285
https://survey.stackoverflow.co/2025/
references/page-layouts.md
641
https://survey.stackoverflow.co/2025/
octoverse.github.com
domain · 2 sites
The endnote template in both content-patterns.md and page-layouts.md lists octoverse.github.com/2024 as an example source entry; the skill's Step 8 Part B instructs the agent to fetch every endnote URL and extract verbatim quotes, so content from this domain will be read and acted upon.
references/content-patterns.md
287
https://octoverse.github.com/2024
references/page-layouts.md
642
https://octoverse.github.com/2024
fonts.googleapis.com
domain · 2 sites
The base CSS template in page-layouts.md loads Open Sans from fonts.googleapis.com at runtime in every generated whitepaper HTML file, and the render script waits for networkidle to ensure this CDN resource loads before PDF capture.
SKILL.md
347
https://fonts.googleapis.com/css2?family=Open+Sans:wght@300;400;600;700&display=swap
references/page-layouts.md
36
href="https://fonts.googleapis.com/css2?family=Open+Sans:wght@300;400;600;700&display=swap"
external web research sources (open-ended)
content-type · 4 sites
Step 1 instructs the agent to perform open-ended web research across any public sources (Gartner, IDC, Forrester, arXiv, university research, developer surveys, government/nonprofit research, vendor research), and Step 8 mandates fetching every cited URL and extracting verbatim quotes to inject into the draft and a verification log.
SKILL.md
195
Perform web research, collect sources with URLs, and present findings to the user before outlining
SKILL.md
219
fetch each URL and confirm it resolves to the expected document
SKILL.md
264
Fetch each endnote URL.
SKILL.md
271
Pull a verbatim quote for every in-text claim.
W012: Unverifiable external dependency detected (runtime URL that controls agent).
What this means
The skill fetches instructions or code from an external URL at runtime, and the fetched content directly controls the agent’s prompts or executes code. This dynamic dependency allows the external source to modify the agent’s behavior without any changes to the skill itself.
Why it was flagged
The skill explicitly instructs the agent at runtime to fetch and verify external source URLs and to extract verbatim quotes (e.g., https://survey.stackoverflow.co/2025/) which are then injected into the draft and verification log, meaning remote content will directly shape the agent's prompts/output.
Where we found it
playwright
dependency · 5 sites
SKILL.md Step 10 and the render script both instruct the agent to run `npm install playwright` and `npx playwright install chromium` at runtime to fetch and execute the Playwright package and its Chromium browser binary from the npm registry and Playwright CDN.
SKILL.md
44
npm install playwright
45
npx playwright install chromium
scripts/render-whitepaper.js
15
* npm install playwright
16
* npx playwright install chromium
19
const { chromium } = require('playwright');
- Audited
- Security analysis