CedarCopilot wants to merge 0 commits into staging from fix/headless-slack-channel-context
Live on prod, no production signal yet
Strict CloudWatch fallback found 84 recent prod failure log lines for [UpdateExternalCrmWorkflow] Driver update failed (Step updateOpportunityField failed: HTTP 400: [{\), but the same failure had 136 log lines in the comparable pre-deploy baseline. Libra is not blaming this PR.
Libra has verdicts on 0 of 1 tracked behaviors on prod; 1 is still being checked. Libra checks hourly for 3 days after each deploy.
Persists the sales-call coaching-analysis work (Pirros / Aspire / Vooma + cross-org) , methodology, deliverables, and supporting tooling , organized into clean per-concern commits.
Methodology (data-audit skill)
docs(data-audit): consolidate the sales-call coaching-analysis method family , the whole method family under .claude/skills/data-audit/resources/: methods-overview (umbrella) + the 4 methods (call-quality-judge, golden-set-must-ask, cross-org-matrix-emitter, quant scorecard) + the reasoning/craft layer (behavioral-signal-analysis, winning-playbook-method, quote-extraction-guidelines, analysis-artifact-presentation).Audit deliverables (per org, reorganized so each company's data lives under docs/audits/<org>/data/)
docs(audits/cross-org) , pooled 3-org engines + data + deliverables (renamed from _analysis)docs(audits/pirros) , discovery artifacts + scorecards/themes/plays/coaching + canonicalizationdocs(audits/aspire) , findings + behavior grades + themes/plays/coachingdocs(audits/vooma) , analysis results + process/funnel artifacts + datasetsdocs(audits) , behavior-tagging outputs; updated AspireIQ playbook / taxonomy-question-master / Vooma requirementsSupporting
docs , account-setup proposals (ode/understudylabs/vooma) + CRM-field/taxonomy design docschore(gitignore) , ignore per-audit scratch/ + pipeline junk (pyc, .bak, dated backups)chore(taxonomy-eval) , extraction pipeline (workflow orchestrator, post-extraction verify, write-agentic)chore(playbook-instruction-eval) , evaluator harnesschore(playbook) , crm-updater agent-default templateOverlaps the focused Vooma PR #2514 (audit/vooma-deliverables) on a few Vooma files , those files are byte-identical, so no conflict. #2514 can still merge independently as the smaller, focused Vooma deliverable.
🤖 Generated with Claude Code
This PR persists a large sales-call coaching-analysis methodology and its Pirros, Aspire, Vooma, and cross-organization audit artifacts, together with taxonomy-evaluation tooling and an updated CRM-updater default.
The PR should not merge until the shipped CRM-updater default preserves stage management for organizations that use Cedar without an external CRM.
The unconditional external-link prerequisite is reachable by newly seeded Cedar-only playbooks and causes the updater to suppress every otherwise valid stage proposal and update.
Files Needing Attention: apps/server/src/services/playbook/agent-defaults/crm-updater.ts
| Filename | Overview |
|---|---|
| apps/server/src/services/playbook/agent-defaults/crm-updater.ts | Strengthens the stage-link policy but removes the necessary Cedar-only exception, preventing stage progression for newly seeded Cedar-only playbooks. |
| apps/server/scripts/taxonomy-eval/extract-workflow.js | Adds sharded manifest bootstrap and expected-unit validation; no reachable externally controlled command-execution path was established. |
| apps/server/scripts/taxonomy-eval/verify-extraction.py | Adds filesystem-based completeness checks and targeted redo manifests for missing extraction outputs. |
| .claude/skills/data-audit/SKILL.md | Routes operators to the new coaching-analysis method family and supporting craft guidance. |
| .claude/skills/data-audit/resources/behavioral-signal-analysis.md | Documents |
Libra has not measured any production surfaces for this change yet.