with one click
session-corpus-audit
Analyze session quality trends — identify high-churn patterns, report waste, flag sessions exceeding 500 tool calls
Menu
Analyze session quality trends — identify high-churn patterns, report waste, flag sessions exceeding 500 tool calls
Adversarially review the last 24h of multi-agent work by combining git history, GitHub issue state, generated analysis artifacts, governance tests, and duplicate-checked follow-up issue creation.
Build self-contained interactive web applications as single HTML files. Use for creating demos, prototypes, interactive tools, and standalone web experiences that work without external servers.
Quick provider/model switching for Hermes CLI — aliases, fallbacks, task routing matrix, and utilization audit pattern.
Quick provider/model switching for Hermes CLI — aliases, fallbacks, task routing matrix, and utilization audit pattern.
Build complex AI systems with declarative programming, optimize prompts automatically, create modular RAG systems and agents with DSPy - Stanford NLP's framework for systematic LM programming
Mandatory planning workflow for ALL GitHub issues — plan, review, approve, then implement.
| name | session-corpus-audit |
| description | Analyze session quality trends — identify high-churn patterns, report waste, flag sessions exceeding 500 tool calls |
| version | 1.0.0 |
| category | coordination |
| tags | ["session","audit","quality","signals"] |
| related_skills | ["session-start-routine","comprehensive-learning"] |
Analyze session signals to identify quality trends and waste patterns.
Session signals live at .claude/state/session-signals/YYYY-MM-DD.jsonl.
Each line is a JSON object with: session_id, transcript_path, cwd, permission_mode, hook_event_name, stop_hook_active, last_assistant_message.
# Last 7 days of session signals
for f in $(ls -t .claude/state/session-signals/*.jsonl | head -7); do
echo "=== $(basename $f) ==="
wc -l "$f"
cat "$f"
done
.claude/state/session-governor/tool-call-count for daily totalsOutput a markdown report with:
When auditing Claude/Codex/Hermes/Gemini sessions for a chatbot or product with existing conversation ratings, load the rated baseline first and treat provider logs as meta-evidence. Do not restart the rubric. Review provider sessions for workflow defects that explain or predict conversation-quality failures: delivery-state overclaims, internal/tool leakage, canary-before-live drift, channel/scope/domain terminology confusion, and user-blame before log inspection.
Use references/conversation-rating-provider-catchup.md for the detailed procedure and output shape.