Add local attempt investigation commands

[Goran Gajic (gorangajic)]

Overview of Changes

When a run attempt fails, there was no way to dig into its trace or logs from the CLI. Agents and people had to open the web app.

• qawolf attempt investigate --attempt-id <id> prints a short summary of what went wrong: failed actions with their errors, failed requests, and recent logs.
• qawolf attempt inspect --attempt-id <id> shows one piece of evidence: an action, a request, a page snapshot (text or HTML), a screenshot saved to a file, console output, or a log record.
• Output is capped in size, and tokens, secrets, cookies and signed URLs are redacted.
• The CLI downloads only the file it needs and never sends the QA Wolf API key to storage URLs.
• The trace reader loads only for these commands, so other commands don't get slower.

Testing

bun run typecheck
bun run lint
bun run format:check
bun run knip
bun run test
bun run build
bun run build:binary

Also checked that the standalone binary loads the trace reader. Not yet tested on Windows.

Checklist

• Changes follow the code style of this project
• Self-review completed
• Tests added/updated (or not applicable)