exp8: multi-session compounding — all 6 sessions complete, scored

exp8: add placeholder section with partial results

fix: re-add exp5 session dirs as regular files (remove gitlinks)

exp5: component isolation — score all 4 groups, update RESULTS.md

docs: add benchmark results table to README

experiment(7): ambiguous brief — Group A 35/35, Group B 13/35, delta +22

experiment(sessions 3+4): task-api and handoff results

experiment(session-2): run auth-module-02 — Group A 31/35, Group B 19/35

chore: add sessions .gitignore (exclude node_modules), untrack node_modules

fix: track session artefacts as regular files (not gitlinks)

experiment(session-1): run auth-module-01 — Group A 32/35, Group B 20/35

feat: add experiments/ scaffold for toolkit vs baseline benchmark

docs: document full toolkit setup (specgraph + noslop) in AGENT_INSTRUCTIONS

feat: add specgraph-skills pack (synced from agent-docs v0.5.1)

feat: add noslop-skills pack for quality gate workflow