email-commslisted

# email-comms > **Runs under AGENTS.md.** These contracts bind without being restated here: Privacy Invariant (`current_base` never outbound), Honesty Firewall, Placeholder/Bracket Ban, Gate Write-back, Domain-Neutral Rule, Browser Automation Contract, Activity Pulse logging, Tracker verify+snapshot, and Sent-Clears-Draft. Inline reminders at point-of-use are intentional; standalone restatements point back to the relevant AGENTS.md section. > **Agent voice.** Read `candidate/modes.yml#agent_voice` (default `standard`) before producing in-chat commentary around drafts. Apply the register from AGENTS.md#mode-switches. The **drafted artifact itself** (email subject + body) is always written in full regardless of register — `writing-style.md` governs its tone. Register governs the **agent commentary**: `exec-summary` = draft only + one action line; `standard` = draft + short bullets on intent/next-step; `technical` = draft + thread-context summary + strategy note; `verbose` = draft + full thread reconstruction + alternatives. ## Inputs - `workspace/tracker.json` — communications[] and applications[] arrays - `workspace/comms/<thread-id>.md` — full raw thread body if previously saved - `candidate/profile.yml` — compensation fields, location, domain - `candidate/targeting.yml` — excluded_companies, cut/keep signals - `candidate/honesty.yml` — claims.do_not_fabricate, confirmed tools/claims - `candidate/evidence.yml` — factual sourcing for all claims - `candidate/writing-style.m