mancitrus/system-prompts-and-models-of-ai-tools
1
0
Fork 0
mirror of https://github.com/x1xhlol/system-prompts-and-models-of-ai-tools.git synced 2026-06-18 07:19:35 +00:00
Code Issues Actions Packages Projects Releases Wiki Activity
35962de933
system-prompts-and-models-o.../salesflow-saas/docs/governance/endpoint-inventory.md
Claude 43058e68cb
feat(dealix): full execution plan + endpoint inventory + customer docs 
FULL_NEXT_STEP_AND_STACK_EXPANSION_AR.md:
  Complete execution plan with 8 gates, 5 phases, stack additions
  (OTel, OIDC, attestations, OpenFGA now; Great Expectations, Unstructured,
  Airbyte next; OPA, Temporal, MCP in radar), backend/frontend upgrades,
  and avoid-now list.

docs/governance/endpoint-inventory.md:
  Trust classification for ALL ~70 endpoints:
  - ~45 Class A (safe auto, read-only)
  - ~15 Class B (approval-gated, side effects)
  - ~6 Class B+ (critical, financial/legal/irreversible)
  - 5 Class C (forbidden)
  With specific trust enforcement requirements per endpoint.

Customer docs:
  ADMIN_SETUP_GUIDE.md — 30-60min setup guide for client IT admin
    (accounts, channels, approvals, data import, compliance, verification)
  EXECUTIVE_QUICKSTART.md — 5-min guide for CEO
    (what you see, 3 daily actions, weekly pack, FAQ)

https://claude.ai/code/session_01W1rJthWDkasijTdXCfxVHs
2026-04-17 06:23:01 +00:00

4.5 KiB
Raw Blame History

Endpoint Inventory — Trust Classification

Parent: MASTER_OPERATING_PROMPT.md
Purpose: Every endpoint classified by risk, side effects, and trust requirements.

Classification Key

Class Meaning Trust Requirements
A Safe auto — read-only, no side effects None
B Approval-gated — causes side effects or external communication correlation_id + approval_token
B+ Critical — financial, legal, or irreversible correlation_id + approval_token + evidence_pack
C Forbidden — never allowed via API Blocked unconditionally

Tier-1 Governance Endpoints

Endpoint Method Class Side Effects Trust Enforced
/executive-room/snapshot GET A None —
/executive-room/weekly-pack GET A None —
/executive-room/risks GET A None —
/executive-room/decisions-pending GET A None —
/executive-room/forecast-vs-actual GET A None —
/approval-center/ GET A None —
/approval-center/stats GET A None —
/approval-center/{id}/approve POST B+ Updates approval status correlation_id via payload
/approval-center/{id}/reject POST B+ Updates approval status correlation_id via payload
/approval-center/{id}/escalate POST B Escalation notification —
/contradictions/ GET A None —
/contradictions/ POST A Creates record —
/contradictions/stats GET A None —
/contradictions/{id}/resolve POST B Status update —
/evidence-packs/assemble POST B Creates SHA256 pack —
/evidence-packs/ GET A None —
/evidence-packs/{id}/review PUT B Status update —
/evidence-packs/{id}/verify GET A None —
/compliance/matrix/ GET A None —
/compliance/matrix/scan POST A Updates control status —
/compliance/matrix/posture GET A None —
/compliance/matrix/risk-heatmap GET A None —
/connectors/governance GET A None —
/connectors/{key}/health-check POST A Updates status —
/model-routing/dashboard GET A None —
/model-routing/health GET A None —
/model-routing/costs GET A None —
/forecast-control/unified GET A None —
/forecast-control/variance GET A None —
/forecast-control/recalibrate POST B Triggers AI reforecast —
/golden-path/run POST B+ Creates approval + evidence correlation_id generated
/golden-path/dossier POST A None (generates schema) —

Core Business Endpoints

Endpoint Method Class Side Effects Trust Required
/leads GET A None —
/leads POST A Creates record —
/leads/import POST B Bulk create —
/deals GET A None —
/deals POST A Creates record —
/deals/{id}/stage PUT B+ Stage change + auto evidence on close Auto evidence on closed_won
/deals/{id} DELETE B Soft delete —

External Communication Endpoints

Endpoint Method Class Side Effects Trust Required
/outreach/* POST B Sends WhatsApp/email/SMS PDPL consent + approval_token
/sequences/* POST B Starts multi-channel sequence PDPL consent + approval_token
/whatsapp-webhook POST A Processes inbound Webhook verification

Strategic Deal Endpoints

Endpoint Method Class Side Effects Trust Required
/strategic-deals/ GET A None —
/strategic-deals/ POST B Creates deal —
/strategic-deals/{id}/negotiate POST B+ Negotiation action correlation_id
/strategic-deals/match POST A AI matching —

Summary

Class Count Enforcement Status
A (safe auto) ~45 No enforcement needed
B (approval-gated) ~15 correlation_id enforced via gateway
B+ (critical) ~6 correlation_id + evidence (golden path enforced)
C (forbidden) 5 Blocked in policy.py