Find out why your AI coding sessions fail — and make your repo ready for the ones that don't.

copied $ pip install agentsesh

v0.10.0 · Python 3.10+ · zero dependencies · GitHub · PyPI

$ sesh audit

Repo Audit ━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━ Metrics ──────────────────────── bootstrap [10/10] ██████████ task_entry_points [ 6/10] ██████░░░░ validation_harness [10/10] ██████████ linting [ 8/10] ████████░░ codebase_map [10/10] ██████████ doc_structure [10/10] ██████████ decision_records [10/10] ██████████ agent_instructions [ 8/10] ████████░░ file_discipline [ 8/10] ████████░░ Score: 89/100 Grade: B

$ sesh analyze

Session Analysis ━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━ Duration: 13 min | 39 tool calls | ~$0.81 Files touched: 12 | Tests: 1 run (1 pass, 0 fail) Grade: A (92/100) What Happened ───────────── 39 tool calls, 2 errors (5% error rate). 3 process issue(s) detected. Tests: 1/1 passed. What To Fix ─────────── [ !!] Reduce error rate (recommended) [ !!] Use dedicated tools instead of Bash (recommended) [ -] Parallelize independent operations (optional)