Add Continue parser

[khadinakbarlabs]

Tool

Continue — open-source AI code assistant for VS Code + JetBrains.

What we'd need

Continue keeps a SQLite history database and per-session JSON. Parse tokens + model only. Never read the prompt or output text.

Likely locations

• ~/.continue/dev_data/ — token usage data
• ~/.continue/sessions/ — per-session JSON
• ~/.continue/index/ — embeddings cache (skip)

Continue's docs explicitly call out ~/.continue/dev_data/<version>/tokens_generated.jsonl — that's almost certainly the right file.

Acceptance

• New entry in the source registry in src/cli.js.
• Returns { source: "continue", model, input_tokens, output_tokens, cost_usd?, synced_at }.
• README + caveats updated.

Tag the PR with parser and mention this issue.