support searching for issues

[schpet]

i am often trying to search issues before creating new ones, this should be easier. some ideas from codex below:

• 1. Add native linear issue search command to the CLI.
     Reason: current workflow requires GraphQL fallback for text search (title/description), which is the most common discovery task and should not require
     query authoring.
  2. Add structured output mode for issue list and issue view that is stable and compact by default (--jsonl).
     Reason: parsing human-formatted CLI output with rg is brittle and misses fields; jsonl would make automation safer and faster.
  3. Add first-class relationship queries (duplicates, related, blocks, blocked by) to issue view.
     Reason: duplicate analysis is central to triage, but currently requires manual inspection or extra API queries.
  4. Improve the skill docs with “preferred query patterns” and ready-to-run snippets for team-scoped full-text search.
     Reason: the skill says “use API as fallback” but doesn’t give a sharp standard workflow for common gaps, so agents re-discover patterns each time.
  5. Add a linear issue find-related <issueId> helper command in the CLI.
     Reason: automatic keyword extraction + team-scoped search + relation output would reduce triage time and improve consistency.

  Best idea: linear issue search as a first-class CLI command.
  Why this is best: it removes the biggest usability gap, reduces dependence on raw GraphQL, benefits both humans and agents immediately, and can internally
  reuse existing API capabilities with low conceptual complexity.