read-web

read-web

Thin adapter for browser-fetch-router read-web. All provider routing, cache, cost, and approval logic live in the CLI — never reimplement here.

Required env

• BFR_AGENT=<agent-name> (e.g. claude, codex, gemini, kimi, opencode, pi)
• BFR_SESSION_ID=<uuid-or-ulid>

Generate BFR_SESSION_ID once per task and reuse it for every CLI call within that task so cost caps, rate limits, and circuit breakers scope correctly.

Command

BFR_AGENT=<agent-name> BFR_SESSION_ID="$session_id" \
  browser-fetch-router read-web <url> --json

Optional flags: --no-cache, --allow-paid, --strict-side-effects, --allow-side-effects, --max-chars N. Side-effects flags (--strict-side-effects, --allow-side-effects) are read-web-specific and govern background writes such as cache persistence; they have no analogue on read-user-tabs or interactive-browser.

Diagnostics

If a CLI call fails with exit code 3 (tool_setup_failed) or 70 (internal_error), run browser-fetch-router doctor --json for setup/health diagnostics, or browser-fetch-router schema --json to inspect the structured-output schema.

Exit codes

Code	Meaning
0	ok
1	content blocked / not found / paywalled / insufficient
2	approval_required or approval_denied
3	tool_setup_failed
4	unsafe_url_blocked
5	cost_cap_exceeded or rate_limited
64	usage_error
70	internal_error

Rules

• Do not retry blocked URLs with the agent's built-in fetch or raw HTTP — they are blocked by policy.
• Do not bypass approval or cost gates.
• Do not store secrets in this skill — pass them via the CLI's documented env vars only.
• Do not switch to read-user-tabs or interactive-browser to fetch a public URL — those modes are higher-trust and reserved for their specific purposes.
• Hosted paid routes require explicit CLI flags such as --allow-paid.