Sort inbound replies into buckets (interested, not_now, ooo, bounced, unsubscribe, wrong_person), match each against cold.md objections, and draft responses. Use when the user says "triage", "sort replies", "what replied", or when cold-report flags items in the needs_review bucket. Reads live replies from FoxReach API. STATUS - stub in v0. Contract is stable; the implementation ships in v0.2. Part of the cold.md suite.
Step 5 of the cold.md suite. Takes raw replies, buckets them, drafts responses that match ## Objections.
This skill defines the contract. The implementation ships in v0.2. Until then, stop and surface to the user:
cold-triage is a stub in v0. Replies live in the FoxReach dashboard at https://foxreach.io/app. For now, paste individual inbounds and the
cold-draftskill will map them against your## Objectionsand draft a reply. Track progress at https://github.com/concaption/cold-md/issues
FOXREACH_API_KEY./cold.md (for the ## Objections section)--campaign-id X, --since N-days, --bucket needs_reviewinterested - positive intent, wants to learn morenot_now - acknowledged, not now, door openooo - out of office / auto-replybounced - deliverability failureunsubscribe - explicit opt-out (STOP processing, mark in FoxReach)wrong_person - "I'm not the right person, contact X instead"needs_review - can't classify confidentlyGET /v1/replies?bucket=needs_review - pull replies## Objections patterns> "objection" in cold.md, use the preferred reply as basePOST /v1/replies/{id}/bucket + draft reply to dashboard for human approvalEvery triaged reply goes to FoxReach's approval queue. Human clicks "send" in the dashboard.