Web Trim

Fetch and summarize a webpage by extracting markdown, removing extraneous content, truncating if needed, and returning a concise summary.

Security Scan

The skill's instructions, requirements, and behavior are consistent with its stated purpose (fetching and summarizing webpages); nothing requested is disproportionate or unrelated.

Purpose & Capability

Name/description (fetch & summarize a webpage) match the runtime instructions: use web_fetch (extractMode: 'markdown'), clean and summarize. The skill declares no credentials, binaries, or installs — all consistent with a simple fetch-and-summarize utility.

Instruction Scope

Instructions are narrowly scoped to spawning a worker session, calling web_fetch with markdown extraction, stripping headers/footers/cookie notices, truncating long content, and returning a concise summary. This is appropriate, though steps like 'truncate to the most relevant sections' are subjective and give the worker some editorial discretion. Also note the skill will fetch arbitrary URLs provided by the user — avoid supplying private/internal links if you don't want their contents retrieved.

Install Mechanism

No install spec and no code files — instruction-only skill. Lowest-risk install posture (nothing is written to disk by the skill itself).

Credentials

No environment variables, credentials, or config paths are requested. The requested capabilities (web_fetch, sessions_spawn) are platform features and align with the stated purpose.

Persistence & Privilege

always:false and no special persistence or config modification. The skill uses worker sessions and returns results; autonomous invocation remains possible (platform default) but is not elevated by the skill itself.

Assessment

This skill appears coherent and limited to fetching and summarizing web pages. Before installing: (1) confirm your platform's web_fetch permission model — the skill will fetch any URL you give it, so don't provide links to private/internal pages or sensitive documents you don't want fetched; (2) test on non-sensitive pages to verify how it truncates and what it strips (it may omit sections you consider important); (3) remember it is instruction-only (no code to audit), so rely on the platform's own sandboxing and audit logs if you need traceability.

Skill: Web-Trim-Summarize

Goal: Fetch a URL, clean the data in a sub-session, and return only the essence.

Protocol:

1. When asked to "Research" or "Summarize" a site:
2. Use sessions_spawn to create a worker.
3. Pass the following instructions to the worker:
   • Use web_fetch on [URL] with extractMode: "markdown".
   • Strip all headers, footers, and cookie notices.
   • If the content is > 10,000 chars, truncate to the most relevant sections.
   • Produce a concise summary (max 500 words).
4. Retrieve the summary via sessions_yield.
5. Paste the final summary into the main chat.