web_fetchlisted

# web_fetch Fetch a URL. ## Operations - `fetch.get(url, headers?)` — returns body + metadata + fetch_hash. - `fetch.json(url)` — parsed JSON. - `fetch.md(url)` — HTML→markdown extraction. - `fetch.paginate(url, pattern)` — follow next links. ## I/O Contract ```yaml io_contract: input: { url, "headers?", "timeout_s?" } output: { status, body, headers, fetch_hash, final_url } idempotent: true side_effects: [network] ``` ## Safety - Robots.txt respected. - Redirects capped at 10. - Body size capped by `acquisition.config.total_wall_budget_s` proxy (configurable). - All fetches logged with URL hashed if `redact_urls: true`. - Bodies stored in MemPalace by `fetch_hash` for audit.