What 1,866 GitHub projects tell their AI coding agents
At a glance
Empirical study: 34% of popular active repos ship agent instruction files (AGENTS.md dominant); half exceed ~100 lines; ~3/5 rules could be script-checked but <1/20 name the check—actionable lint: short map, name enforcers, one rulebook. This workflow is listed under learnable-paradigm, agents-md, coding-agents, research, context.
Agents / tools
AGENTS.md, CLAUDE.md, Codex, Claude Code, GitHub Copilot
AWS Labs open-source (5k+★): one harness-neutral core turns Claude Code/Codex/Cursor/Copilot/Kiro/opencode into structured, verifiable delivery workflows via `aidlc config --harness` + doctor—install script, not a new agent.
MCP-connected runtime QA maps critical flows and entry points, then re-tests the blast radius of each Claude Code/Cursor/Codex change against live Postgres/Kafka/APIs/MCP tools—contracts, authz, OWASP, errors—before merge, with self-heal when behavior changes on purpose.
SuperCompress v2 compresses context against the query (keeps answer-critical lines verbatim): ~64% mean cut on their B5 coding-agent suite with 24/24 evidence-retention passes; MCP/plugin for Cursor/Claude Code/Codex plus hosted API and open weights.
GitLab 19.5 makes the MCP server GA on Free/Premium/Ultimate: Claude Code, Cursor, Codex and other MCP clients get consolidated tools (get_merge_request, get_pipeline) plus Duo Agent Platform flow handoff under existing GitLab permissions, with GitLab.com rate limits of 60/min Free and 600/min paid.