research-complisted

# research-comp Benchmark market compensation for a role+location via web search. Produces a cited `workspace/research/comp-bench-<role-slug>-<loc-slug>-<yyyy-mm>.md` artifact, emits a COMP BENCHMARK addendum block (when invoked from `evaluate-job`), and proposes — never auto-applies — gate write-backs when market data shows candidate targets are misaligned. > **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 the comp benchmark summary. Apply the register from AGENTS.md#mode-switches. The benchmark artifact is always written in full — register governs the in-chat presentation: `exec-summary` = floor/midpoint/ceiling numbers + 1-line confidence note; `standard` = scannable table of sources + band + alignment note; `technical` = per-source analysis + confidence rationale; `verbose` = full artifact + sourcing discussion. --- ## Inputs | Source | Fields read | | --- | --- | | `candidate/profile.yml` | `compensation.minimum_base`, `compensation.target_base`, `compensation.expected_base`, `compensation.oe_min_