Kindo × Deloitte — Sprint & Release Plan
A.1–A.5 Production Readiness · June 9–22, 2026
Definition of Done: Production-ready releases with P1/P2 bugs remediated, ready for handoff to Krishna's D&RaaS team
Sprint 1Jun 9–11
Sprint 2Jun 11–13
Sprint 3Jun 13–15
Sprint 4Jun 16–18
Sprint 5Jun 18–20
Sprint 6Jun 20–22
Goal: Krishna's analysts run all 5 core agents in production — threat monitoring, intel enrichment, hunting, detection engineering, and exposure assessment — end-to-end without data gaps or tool failures.
Sprint 1 — Unblock Critical Paths
Jun 9–11
🎯 Eliminate the 3 urgent blockers that prevent A.1 and A.2 from functioning
-
URGENT
TEK-154 Swimlane 0-record fetch — searches returning zero kills A.1 monitoring workflows A.1
-
IN PROG
TEK-95 VirusTotal threat-intel — no enrichment = A.2 is blind. Oldest urgent (Apr 11) A.2
-
IN PROG
TEK-106 ThreatConnect reliability — high failure rate breaks intel correlation A.2
-
READY
TEK-58 SailPoint ISC → push to Deloitte prod validation A.1
-
READY
TEK-141 Swimlane record update → push to Deloitte validation A.1
Sprint 2 — Resolve Blocked + Complete Reviews
Jun 11–13
🎯 Unblock SailPoint writes, land Jira capabilities, complete SQL MCP for A.3/A.5 data access
-
BLOCKED
TEK-139 SailPoint write ops — escalate for failing-call evidence from Deloitte A.1 A.4
-
REVIEW
APO-117 Jira upload attachments — evidence back into tickets A.3 A.4
-
REVIEW
APO-95 SQL Database MCP — data store queries for hunting + exposure A.3 A.5
-
ADDED
APO-131 Jira priority default bug — fix medium-priority default A.4
-
ADDED
TEK-155 SailPoint operational tool policies — production policy coverage A.1
-
IN PROG
TEK-128 Saviynt connector — cloud IAM for exposure assessment A.5
Sprint 3 — Validate + Release Prep
Jun 13–15
🎯 P1/P2 bug sweep, Deloitte validation sign-off, integration testing across all 5 agents
-
P1/P2
Bug sweep across A.1–A.5 agent workflows — remediate all P1 and P2 issues
-
VALIDATE
Deloitte validation of TEK-58 (SailPoint ISC) + TEK-141 (Swimlane) in production
-
HIGH
Structured JSON — agent output schemas for downstream consumption ALL
-
85%
Long-running reliability hardening — ship remaining 15% ALL
-
E2E
Integration testing: each agent end-to-end against Deloitte test systems
-
HANDOFF
Release 1 package to Krishna's team with release notes + known issues
Goal: Operations leads can monitor agent performance from Command Center, detection rules connect to GitHub, multi-agent workflows are production-stable, and the platform is client-deployment ready.
Sprint 4 — A.4/A.5 Enhancement + Dev Tooling
Jun 16–18
🎯 Complete GitHub connectivity for detection eng, land Python MCP, finish Saviynt
-
TO DO
TEK-143 GitHub OAuth on SMK — detection rules live in GitHub, no OAuth = no connection A.4
-
TO DO
APO-119 Python MCP server support — custom integration extensibility ALL
-
IN PROG
TEK-128 Saviynt connector completion — if not landed in Sprint 2 A.5
-
TO DO
Step progress indicator in agent run chat — UX for long-running workflows
-
HIGH
Show relevant tools per prompt — reduce noise in agent tool selection
Sprint 5 — Command Center + Observability
Jun 18–20
🎯 Ops leads can track agent cost, performance, and failures from one dashboard
-
TO DO
Token logging — per-agent token logging & cost tracking CMD CTR
-
TO DO
Usage export — data export beyond 60 days CMD CTR
-
ADDED
Run ended column in agent runs tab CMD CTR
-
HIGH
Failure logs — export failure logs from failed run OPS
-
HIGH
Error messages — meaningful actionable error messages UX
-
25%
SMK Observability — telemetry, log forwarding, dashboards INFRA
Sprint 6 — Multi-Agent + Release Polish
Jun 20–22
🎯 Multi-agent orchestration stable, whitelabel ready, full regression across A.1–A.5
-
61%
Multi-Agent Orch — agent-to-agent workflows and spawn controls ALL
-
REVIEW
Whitelabel flag config schema for custom deploys DEPLOY
-
TO DO
Resizable chat — input & output windows UX
-
IN PROG
SMK Release Process — upgrade and backport hardening INFRA
-
REGRESSION
Full regression test across A.1–A.5 with Release 1 + Release 2 changes
-
HANDOFF
Release 2 package to Krishna's team — platform hardened, ops tooling live
Critical Path by Sprint
🔴 Sprint 1 Critical Path (Jun 9–11)
- TEK-154 Swimlane 0-record fetch → without this fix, A.1 Threat Monitoring cannot reliably query case data → all downstream monitoring workflows fail → Release 1 ships with a broken lead agent
- TEK-95 VirusTotal + TEK-106 ThreatConnect → these two run in parallel but BOTH must land → A.2 Threat Intel has zero automated enrichment without them → analysts fall back to manual lookups
- Gate: TEK-154 + (TEK-95 ∥ TEK-106) must be code-complete by Jun 11 or Release 1 date is at risk
🟡 Sprint 2 Critical Path (Jun 11–13)
- TEK-139 SailPoint write ops → BLOCKED on Deloitte evidence → if not unblocked by Jun 11, A.1/A.4 ship read-only → remediation actions require manual SailPoint work
- APO-95 SQL Database MCP → gate for both A.3 (hunt queries) and A.5 (exposure data) → must clear review by Jun 12 or two agents lose data access capabilities
- TEK-128 Saviynt → sole cloud IAM connector for A.5 CTEM → if it slips, A.5 launches without IAM exposure visibility
- Gate: APO-95 review complete + TEK-139 escalation response received by Jun 12
🔵 Sprint 3 Critical Path (Jun 13–15)
- Deloitte validation of TEK-58 + TEK-141 → both are READY but need Deloitte sign-off to go live → if Harshal/team doesn't validate by Jun 14, SailPoint ISC and Swimlane updates don't make R1
- P1/P2 bug sweep → must run against ALL 5 agents in Deloitte test environment → requires ENG-8425 SAP connectivity to be stable (currently in progress)
- Long-running reliability (85%→ship) → remaining 15% is the difference between agents timing out on complex workflows vs completing them
- Gate: Krishna's team available for acceptance testing Jun 14–15. If not scheduled by Jun 12, handoff slips.
🟣 Sprint 4 Critical Path (Jun 16–18)
- TEK-143 GitHub OAuth on SMK → detection rules live in GitHub → without OAuth, A.4 Detection Eng can't pull or push rules → agent is functional but disconnected from rule source-of-truth
- APO-119 Python MCP → enables custom integration patterns → prerequisite for any bespoke connector work in Sprint 5–6
- Gate: GitHub OAuth connection config available from Divya Singhal by Jun 16
🟠 Sprint 5 Critical Path (Jun 18–20)
- Per-agent token logging → Command Center metrics depend on this → without it, org-wide usage graphs and cost tracking have no data source
- Export failure logs → ops team can't diagnose agent failures at scale without this → production support readiness depends on it
- Gate: Token logging pipeline instrumented by Jun 19 or Command Center maturity items cascade into post-R2
🟢 Sprint 6 Critical Path (Jun 20–22)
- Multi-Agent Orchestration (61%→ship) → agent-to-agent handoffs are the foundation for complex Deloitte workflows → remaining 39% determines whether multi-step investigation chains work
- Full regression → must cover R1 + R2 changes together → any regressions introduced by Sprint 4–5 work caught here
- Gate: Regression pass + Krishna team sign-off by Jun 22 EOD
Agent × Integration Dependency Map
A.1 Threat Monitoring
TEK-154 Swimlane 0-fetch ⚠️
TEK-139 SailPoint write 🔒
TEK-58 SailPoint ISC ✅
TEK-141 Swimlane update ✅
TEK-155 SailPoint policies
A.2 Threat Intel
TEK-95 VirusTotal ⚠️
TEK-106 ThreatConnect ⚠️
A.3 Threat Hunt
APO-111 Jira read ✅
APO-121 Jira fields ✅
APO-117 Jira upload 🔍
APO-95 SQL MCP 🔍
A.4 Detection Eng
APO-117 Jira upload 🔍
APO-131 Jira priority bug
TEK-143 GitHub OAuth
A.5 CTEM
TEK-128 Saviynt 🔨
TEK-91 Okta writes 📋
APO-95 SQL MCP 🔍
Resource Constraints, Blockers & Mitigation
🚫 Blockers (External Dependencies)
- TEK-139 SailPoint writes — Blocked on Deloitte providing failing-call evidence. Mitigation: Escalate directly to Harshal; offer to pair on evidence capture this week.
- APO-42/44 SAP JCo/RFC — Need JCo params + cert config from Deloitte. Mitigation: Out of R1 scope; park for R2+ unless params arrive early.
- TEK-97 MS Exchange — Needs licensed M365 mailbox. Mitigation: Deloitte infra team dependency; out of R1/R2 scope.
- TEK-91 Okta writes — Deloitte-side enablement. Mitigation: Request confirmation by Sprint 4; plan A.5 workaround without Okta write.
⚡ Assumptions
- Kindo engineering capacity sustains current velocity on TEK-95, TEK-106, TEK-154 through June 15
- Deloitte validation turnaround for READY items (TEK-58, TEK-141) is ≤3 business days
- Krishna's team is available for Release 1 acceptance testing June 14–15
- Structured JSON output spec can be defined from existing agent schemas (no new product design needed)
- ENG-8425 SAP test-system connectivity does not block A.1–A.5 (separate workstream)
👥 Resource Constraints
- Kindo eng bandwidth: 6 items IN PROGRESS simultaneously — any new urgent pull risks slippage on TEK-95/TEK-106
- Deloitte validation: Harshal Sarode is the requester on 8+ items — single point of contact bottleneck
- Kindo UX session (Jun 9): Product team time diverted to design alignment — protect eng focus on R1 blockers
- T&C (Tony/Joana): Escalation + stakeholder alignment capacity — not implementation resources
🛡️ Mitigation Plans
- If TEK-154 slips past Sprint 1: A.1 ships with known limitation + monitoring workaround documented for Krishna's team
- If TEK-95/106 don't land by June 13: A.2 ships with manual enrichment fallback; automated enrichment becomes R2 Sprint 4 priority
- If TEK-139 stays blocked: A.1/A.4 ship read-only for SailPoint; write-back becomes R2 deliverable
- If structured JSON isn't feasible in Sprint 3: Move to Sprint 4; R1 ships with current output format + schema spec doc for R2