Sprints & cards

The execution domain. Once a spec is validated, work splits into sprints (logical delivery slices) and cards (single execution units). Pulse exposes 26 of its 216 MCP tools for this stage:

Sprint lifecycle (8) — create / update / move / get / get_context / list, AI-assisted suggest_sprints, and assign_tasks_to_sprint.
Sprint evaluation (4) — submit / list / get / delete sprint evaluation (the review → closed gate).
Card lifecycle (7) — create / get / get_task_context / get_task_conclusions / update / move / delete.
Card dependencies (3) — add_card_dependency / remove_card_dependency / get_card_dependencies.
Listing, blockers, analytics (4) — list_cards_by_status, list_blockers, get_traceability_report, get_analytics.

This page covers sprint scoping, sprint quality evaluation, and the card lifecycle. The task validation gate (submit_task_validation) and the test theater anti-pattern (update_test_scenario_status) live on the Validation gates page. Card Q&A and comments live on Comments & questions; card knowledge and attachments on Knowledge & attachments; copy-from-spec mockups and architecture on Architecture & mockups. Source-of-truth citations: okto-pulse-feature-inventory.md:256–271 (card lifecycle), :457–475 (sprint tools), :395–410 (analytics + blockers + traceability), :967–981 (status enums), :982–1021 (governance gates).

Sprint vs. card

	Sprint	Card
What it is	A logical delivery slice — scope + objective + expected outcome.	A single execution unit — one implementer’s piece of the slice.
Created from	A `validated` spec (typically via `suggest_sprints`).	A spec, a sprint, or directly on a board. Type is `normal` / `test` / `bug`.
Statuses	`draft → active → review → closed`	`not_started → started → in_progress → validation → done`; with `on_hold` / `cancelled` side states.
Hard gate	`submit_sprint_evaluation` before `review → closed`.	`submit_task_validation` (independent reviewer) + conclusion + completeness + drift before `validation → done`.

A sprint is not a time-box. It is a scope contract: “these N cards go together; close the sprint when they are validated.” Sprints scope a release more than a calendar.

Status flow and gates

Stage	Transition	MCP tool	Gate
Sprint	`draft → active`	`okto_pulse_move_sprint`	none
Sprint	`active → review`	`okto_pulse_move_sprint`	none
Sprint	`review → closed`	`okto_pulse_move_sprint`	`submit_sprint_evaluation` recorded; passed test scenarios re-validated
Card	`not_started → started`	`okto_pulse_move_card`	none
Card	`started → in_progress`	`okto_pulse_move_card`	bug cards: at least one `failing` test scenario must exist (test-first gate)
Card	`in_progress → validation`	`okto_pulse_move_card`	none
Card	`validation → done`	`okto_pulse_move_card`	conclusion + completeness + drift recorded; if Task Validation Gate enabled, an independent reviewer must `submit_task_validation` — see Validation gates

Sprint close re-validates every test scenario marked passed on cards in the sprint. A scenario marked passed without the structured evidence recorded by update_test_scenario_status is treated as theater — sprint close fails and the scenario regresses to pending until evidence is supplied. See the Validation gates page for the evidence schema.

Sprint tool index

Lifecycle (6)

create_sprint, update_sprint, move_sprint, get_sprint, get_sprint_context, list_sprints.

Suggest & assign (2)

suggest_sprints (AI-assisted slicing), assign_tasks_to_sprint.

Evaluation (4)

submit_sprint_evaluation, list_sprint_evaluations, get_sprint_evaluation, delete_sprint_evaluation.

Q&A (3)

ask_sprint_question, answer_sprint_question, delete_sprint_question. See Comments & questions for full Q&A coverage across stages.

Sprint lifecycle

`okto_pulse_create_sprint`

Source: server.py:11272. Create a sprint under a spec. Most agents use suggest_sprints first and create from the suggestion’s payload.

spec_id:          "spc_b921..."
title:            "Sprint 1 — middleware + 429 path"
objective:        "Wire Redis token-bucket middleware with fail-open and 429 emission."
expected_outcome: "Burst limits enforced on /search; integration tests cover AC1, AC2, AC3."

`okto_pulse_update_sprint`

Source: server.py:11345. Update sprint fields.

`okto_pulse_move_sprint`

Source: server.py:11426. Advance status. The review → closed gate enforces sprint evaluation and test-scenario evidence recheck.

`okto_pulse_get_sprint`

Source: server.py:11472. Get a sprint by ID.

`okto_pulse_get_sprint_context`

Source: server.py:11520. Rich bundle: sprint + parent spec + all assigned cards + KG context. Use this before activating or closing a sprint.

`okto_pulse_list_sprints`

Source: server.py:11646. List sprints on a board or spec.

`okto_pulse_assign_tasks_to_sprint`

Source: server.py:11682. Assign one or more cards to a sprint. Cards can move between sprints (e.g. when a sprint slips, deferred cards roll forward).

sprint_id: "spr_a01b..."
card_ids:  ["card_771a...", "card_88b3...", "card_ad14..."]

`okto_pulse_suggest_sprints`

Source: server.py:12128. AI-assisted: slice a validated spec into one or more reviewable deliveries. Returns sprint candidates with title, objective, expected outcome, and recommended card list.

spec_id: "spc_b921..."

Sprint evaluation

The sprint evaluation is the quality gate for review → closed. It captures a written retrospective with structured ratings (a sprint cannot close silently).

Tool	Line	Description
`okto_pulse_submit_sprint_evaluation`	`server.py:11724`	Submit an evaluation (rating + narrative)
`okto_pulse_list_sprint_evaluations`	`server.py:11799`	List evaluations recorded on a sprint
`okto_pulse_get_sprint_evaluation`	`server.py:11831`	Get one evaluation by ID
`okto_pulse_delete_sprint_evaluation`	`server.py:11861`	Delete an evaluation

sprint_id: "spr_a01b..."
rating:    "good"
narrative: "All 3 cards closed; AC1-AC3 fully covered. One card slipped to Sprint 2 due to Redis cluster auth issue (linked decision dec_e7ca already updated). No drift on the validated cards."

Card tool index

Lifecycle (5)

create_card, get_card, update_card, move_card, delete_card.

Context bundles (2)

get_task_context (rich pre-implementation bundle), get_task_conclusions (post-close summary).

Dependencies (3)

add_card_dependency, remove_card_dependency, get_card_dependencies.

Listing & analytics (4)

list_cards_by_status, list_blockers, get_traceability_report, get_analytics.

Card lifecycle

`okto_pulse_create_card`

Source: server.py:1550. Create a card. Card type controls downstream gates: normal (default), test (test scaffolding), bug (test-first gate).

board_id:  "brd_abc123"
spec_id:   "spc_b921..."
sprint_id: "spr_a01b..."
type:      "normal"
priority:  "high"
title:     "Implement token-bucket Redis primitive"
description: "Implement INCRBYFLOAT + EXPIRE primitive in middleware/rate_limit.py. Cover AC1 + BR br_3da7."

`okto_pulse_get_card`

Source: server.py:1799. Get full card details.

`okto_pulse_get_task_context`

Source: server.py:1880. The single most-called tool inside an implementation session. Returns card + spec + business rules + test scenarios + decisions + KG context — everything an implementer agent needs to start work without re-reading the spec from scratch.

Always call get_task_context before moving a card to in_progress. The KG context block surfaces decisions another agent recorded on a similar card; reading them often cuts implementation time substantially.

`okto_pulse_get_task_conclusions`

Source: server.py:2090. Get conclusion text + completeness score (0–100) + drift score (0–100) for a card that has reached validation or done.

`okto_pulse_update_card`

Source: server.py:2145. Update card fields, including conclusion, completeness, drift, and bug_severity. The conclusion fields are required before move_card can transition validation → done.

card_id:      "card_771a..."
conclusion:   "Token-bucket implemented with INCRBYFLOAT + EXPIRE. Atomic via Lua script (RT_LIMITER_INCR.lua). Fail-open path tested via Redis-down chaos test (test_rate_limit_fail_open)."
completeness: 95
completeness_justification: "All AC1 paths covered; one perf test (sustained 5 req/s for 10 minutes) deferred — not blocking, tracked via card_88b3 follow-up."
drift:        10
drift_justification: "Spec called for INCRBY; switched to INCRBYFLOAT to handle fractional refill rate (decision dec_f201 added)."

`okto_pulse_move_card`

Source: server.py:2291. Advance card status. Forward gates are server-enforced.

`okto_pulse_delete_card`

Source: server.py:2396. Permanently delete a card. Use archive_tree (see Architecture & mockups) for soft-delete.

Dependencies

Tool	Line	Description
`okto_pulse_add_card_dependency`	`server.py:2440`	Add a `depends_on` relationship between two cards
`okto_pulse_remove_card_dependency`	`server.py:2477`	Remove a dependency
`okto_pulse_get_card_dependencies`	`server.py:2503`	List blockers and dependents for a card

card_id:        "card_88b3..."
depends_on_id:  "card_771a..."

A card whose depends_on set contains an unfinished card is reported in list_blockers. The card itself can still move forward if its dependencies are satisfied — this is hard data, not a hard gate.

Listing, blockers, traceability, analytics

These four tools span sprints, cards, and the spec contract — they answer board-level health questions and are commonly the first calls a planning agent makes.

`okto_pulse_list_cards_by_status`

Source: server.py:2544. List cards filtered by status, type, sprint, or spec.

`okto_pulse_list_blockers`

Source: server.py:8199. List cards currently blocked — status=on_hold or with unresolved dependencies. Pulse classifies both shapes the same way; an agent reviewing blockers does not need to query both lists.

`okto_pulse_get_traceability_report`

Source: server.py:10713. Full lineage: spec → BRs → TRs → scenarios → cards → validations. The single best call for “what evidence have we produced for this spec?” — runs entirely inside Pulse, no KG round-trip needed.

spec_id: "spc_b921..."

`okto_pulse_get_analytics`

Source: server.py:7893. Board-level metrics: velocity (cards/sprint), quality (validation pass rate, drift average), coverage (AC → scenario %), and per-agent breakdowns.

Permissions

Action	Permission flag
Create / update sprint	`Permissions.SPRINTS_CREATE`, `Permissions.SPRINTS_UPDATE`
Move sprint forward / backward	`Permissions.SPRINTS_MOVE`, `Permissions.SPRINTS_MOVE_BACKWARD`
Submit sprint evaluation	`Permissions.SPRINTS_EVALUATE`
Create / update card	`Permissions.CARDS_CREATE`, `Permissions.CARDS_UPDATE`
Move card forward / backward	`Permissions.CARDS_MOVE`, `Permissions.CARDS_MOVE_BACKWARD`
Add / remove card dependency	`Permissions.CARDS_DEPEND`
Delete sprint / card	`Permissions.SPRINTS_DELETE`, `Permissions.CARDS_DELETE`
Read traceability / analytics	`Permissions.BOARD_READ`

For task validation gate permissions (Permissions.VALIDATOR_*), see Validation gates.

Errors

{
  "error": "CONFLICT",
  "message": "Cannot move bug card to 'in_progress' — no failing test scenario linked. Add a failing scenario before starting bug work (test-first gate).",
  "detail": {"card_id": "card_bug12...", "card_type": "bug", "linked_scenarios_failing": 0}
}

Code	Most common cause
`VALIDATION_ERROR`	Conclusion / completeness / drift fields empty when moving `validation → done`; sprint evaluation submitted without `narrative`.
`NOT_FOUND`	Card / sprint / dependency ID belongs to another board.
`FORBIDDEN`	Agent’s preset does not include the required permission flag.
`CONFLICT`	Bug card → `in_progress` without a failing test scenario; sprint → `closed` without an evaluation; card → `done` without conclusion fields or task validation.

Next steps

Validation gates

submit_task_validation, update_test_scenario_status (anti-theater evidence), and the Validator preset.

Specs

Author the contract a sprint and its cards execute against.

Comments & questions

Card Q&A and free-form comments / choice polls.

ADLC pipeline

Where sprints and cards sit in the end-to-end flow.

​Sprints & cards

​Sprint vs. card

​Status flow and gates

​Sprint tool index

Lifecycle (6)

Suggest & assign (2)

Evaluation (4)

Q&A (3)

​Sprint lifecycle

​okto_pulse_create_sprint

​okto_pulse_update_sprint

​okto_pulse_move_sprint

​okto_pulse_get_sprint

​okto_pulse_get_sprint_context

​okto_pulse_list_sprints

​okto_pulse_assign_tasks_to_sprint

​okto_pulse_suggest_sprints

​Sprint evaluation

​Card tool index

Lifecycle (5)

Context bundles (2)

Dependencies (3)

Listing & analytics (4)

​Card lifecycle

​okto_pulse_create_card

​okto_pulse_get_card

​okto_pulse_get_task_context

​okto_pulse_get_task_conclusions

​okto_pulse_update_card

​okto_pulse_move_card

​okto_pulse_delete_card

​Dependencies

​Listing, blockers, traceability, analytics

​okto_pulse_list_cards_by_status

​okto_pulse_list_blockers

​okto_pulse_get_traceability_report

​okto_pulse_get_analytics

​Permissions

​Errors

​Next steps

Validation gates

Specs

Comments & questions

ADLC pipeline

Sprints & cards

Sprint vs. card

Status flow and gates

Sprint tool index

Sprint lifecycle

`okto_pulse_create_sprint`

`okto_pulse_update_sprint`

`okto_pulse_move_sprint`

`okto_pulse_get_sprint`

`okto_pulse_get_sprint_context`

`okto_pulse_list_sprints`

`okto_pulse_assign_tasks_to_sprint`

`okto_pulse_suggest_sprints`

Sprint evaluation

Card tool index

Card lifecycle

`okto_pulse_create_card`

`okto_pulse_get_card`

`okto_pulse_get_task_context`

`okto_pulse_get_task_conclusions`

`okto_pulse_update_card`

`okto_pulse_move_card`

`okto_pulse_delete_card`

Dependencies

Listing, blockers, traceability, analytics

`okto_pulse_list_cards_by_status`

`okto_pulse_list_blockers`

`okto_pulse_get_traceability_report`

`okto_pulse_get_analytics`

Permissions

Errors

Next steps