Linearis CLI Reference

Verified against Linearis v2026.4.9 (2026-05-31). ⚠️ READ vs WRITE. Linear READS → the local replica by direct SQL, or linear_read_ticket <ID>. Never shell linearis issues read for a routine read — it 429s the shared fleet quota. WRITES → linearis. Read Gotchas before scripting.

Setup check (first, every session)

node "${CLAUDE_SKILL_DIR}/scripts/identity-report.mjs" — one line per identity (tenant, human, team, cloud host), and every unresolved line is a stop-and-say (CTL-2300). Paths like that name files inside this skill: Claude Code fills in ${CLAUDE_SKILL_DIR}; on another harness set CLAUDE_SKILL_DIR to this SKILL.md's directory, or stop and report skill_dir_unresolved. A write addressed to the wrong team or the wrong workspace does not error; it lands somewhere plausible, which is why this runs before the first read as well as the first write.

Reading Linear

Single source of the Linear read rule — other skills point here, they don't restate it.