feat(quick-dev): add previous story continuity to context loading

[alexeyv]

Summary

• When quick-dev infers the intent is an epic story, it now scans {implementation_artifacts} for the most recent completed spec from the same epic
• Loads the previous spec and extracts Code Map, Design Notes, Spec Change Log, and task list as continuity context for step-02 planning
• Falls back to asking the user about in-review specs if no done spec exists; non-epic intents skip silently

Stacked on #2185.

Test plan

• Quality gate passes (npm ci && npm run quality)
• Run quick-dev with an epic story intent — verify it loads the previous completed spec
• Run quick-dev with a non-epic intent — verify no previous story detection occurs
• Run quick-dev on story 1 of an epic — verify silent skip when no prior spec exists

🤖 Generated with Claude Code

[coderabbitai]

📝 Walkthrough

Walkthrough

Adds a "Previous story continuity" context-loading step during step 1's clarification phase. When the current intent matches an epic story, the workflow searches for earlier status: done specs from the same epic with lower story numbers, extracts continuity fields (Code Map, Design Notes, Spec Change Log, task list) from the latest match, and loads them downstream. If no done spec exists but an in-review spec does, the user is prompted to confirm loading it.

Changes

Cohort / File(s)	Summary
Step 01 Context Loading src/bmm-skills/4-implementation/bmad-quick-dev/step-01-clarify-and-route.md	Added "Previous story continuity" context-loading logic that detects epic stories, locates earlier completed or in-review specs from the same epic, and extracts continuity fields for downstream planning steps.

Estimated code review effort

🎯 1 (Trivial) | ⏱️ ~3 minutes

Possibly related PRs

• feat(quick-dev): add planning artifact awareness for context-informed specs #2185: Modifies step-01's context-loading behavior by adding planning-artifact awareness, complementing the new prior-epic/story continuity load.
• feat(quick-dev): one-shot route generates spec trace file #2121: Alters step-01's routing and clarification logic in the same file, specifically slug/spec_file selection, related to the continuity loading enhancement.
• refactor(quick-dev): replace unconditional artifact scan with intent cascade #2105: Restructures artifact-scanning into a 3-tier intent cascade in step-01, which intersects with the new continuity context-loading integration point.

🚥 Pre-merge checks | ✅ 3

✅ Passed checks (3 passed)

Check name	Status	Explanation
Title check	✅ Passed	The title 'feat(quick-dev): add previous story continuity to context loading' accurately summarizes the main change—adding previous story continuity to the context loading process in quick-dev.
Description check	✅ Passed	The description is clearly related to the changeset, explaining what the feature does, how it works, and providing a test plan aligned with the actual changes documented in the raw summary.
Docstring Coverage	✅ Passed	No functions found in the changed files to evaluate docstring coverage. Skipping docstring coverage check.

_{✏️ Tip: You can configure your own custom pre-merge checks in the settings.}

✨ Finishing Touches

🧪 Generate unit tests (beta)

• Create PR with unit tests
• Commit unit tests in branch feat/quick-dev-previous-story-awareness

---

Thanks for using CodeRabbit! It's free for OSS, and your support helps us grow. If you like it, consider giving us a shout-out.

❤️ Share

• X
• Mastodon
• Reddit
• LinkedIn

_{Comment @coderabbitai help to get the list of available commands and usage tips.}

[coderabbitai]

🧹 Nitpick comments (2)

src/bmm-skills/4-implementation/bmad-quick-dev/step-01-clarify-and-route.md (2)

56-56: Make the task extraction target explicit to match template structure.

“task list” is vague. Please name the exact source section (e.g., Tasks & Acceptance → Execution checklist) so extraction is deterministic.

🤖 Prompt for AI Agents

Verify each finding against the current code and only fix it if needed.

In `@src/bmm-skills/4-implementation/bmad-quick-dev/step-01-clarify-and-route.md`
at line 56, Replace the vague phrase "task list" with an explicit source section
name so extraction is deterministic; update the sentence that lists extracted
items to reference the template section "Tasks & Acceptance → Execution
checklist" (or the exact section title used in your template) instead of "task
list" so downstream step-02 planning knows the precise extraction target; ensure
the updated phrase appears in the line containing "Extract its **Code Map**,
**Design Notes**, **Spec Change Log**, and **task list**" and matches the
template's section header string exactly.

---

55-55: Clarify “most recent” vs “highest story number” to avoid ambiguous behavior.

Line 55 conflates chronological recency with ordinal predecessor. If you intend numeric predecessor, say that explicitly (e.g., “select the highest lower story number”), or define tie-breakers by completion timestamp if true recency is intended.

🤖 Prompt for AI Agents

Verify each finding against the current code and only fix it if needed.

In `@src/bmm-skills/4-implementation/bmad-quick-dev/step-01-clarify-and-route.md`
at line 55, The phrase "Load the most recent one (highest story number below
current)" is ambiguous between chronological recency and numeric predecessor;
update the instruction text in step 3 to explicitly state the intended selection
rule—either change it to "select the highest lower story number" if you mean the
numeric predecessor (i.e., the largest story_number < current story_number), or
state "select the most recently completed story among those with story_number <
current" and add a tie-breaker by completion timestamp if true recency is
intended; ensure you edit the exact sentence "Load the most recent one (highest
story number below current)" so it unambiguously reflects the chosen rule.

🤖 Prompt for all review comments with AI agents

Verify each finding against the current code and only fix it if needed.

Nitpick comments:
In `@src/bmm-skills/4-implementation/bmad-quick-dev/step-01-clarify-and-route.md`:
- Line 56: Replace the vague phrase "task list" with an explicit source section
name so extraction is deterministic; update the sentence that lists extracted
items to reference the template section "Tasks & Acceptance → Execution
checklist" (or the exact section title used in your template) instead of "task
list" so downstream step-02 planning knows the precise extraction target; ensure
the updated phrase appears in the line containing "Extract its **Code Map**,
**Design Notes**, **Spec Change Log**, and **task list**" and matches the
template's section header string exactly.
- Line 55: The phrase "Load the most recent one (highest story number below
current)" is ambiguous between chronological recency and numeric predecessor;
update the instruction text in step 3 to explicitly state the intended selection
rule—either change it to "select the highest lower story number" if you mean the
numeric predecessor (i.e., the largest story_number < current story_number), or
state "select the most recently completed story among those with story_number <
current" and add a tie-breaker by completion timestamp if true recency is
intended; ensure you edit the exact sentence "Load the most recent one (highest
story number below current)" so it unambiguously reflects the chosen rule.

---

ℹ️ Review info

⚙️ Run configuration

Configuration used: Path: .coderabbit.yaml

Review profile: CHILL

Plan: Pro

Run ID: ed73f814-58b6-4d47-bb4d-172f0fa8093b

📥 Commits

Reviewing files that changed from the base of the PR and between d51e215 and 7bedd8d.

📒 Files selected for processing (1)

• src/bmm-skills/4-implementation/bmad-quick-dev/step-01-clarify-and-route.md

[augmentcode]

🤖 Augment PR Summary

Summary: Enhances quick-dev’s context loading to carry forward continuity when the current intent appears to be a story within an epic.

Changes:

• When an epic-story intent is detected, scans {implementation_artifacts} for the most recent prior spec in the same epic with status: done (or optionally in-review)
• Loads key sections (Code Map, Design Notes, Spec Change Log, task list) from that previous spec to inform step-02 planning

_{🤖 Was this summary useful? React with 👍 or 👎}

[augmentcode]

Review completed. 2 suggestions posted.

Comment augment review to trigger a new review at any time.

[alexeyv]

Triage complete: 2 findings — FIX: 0, DISMISS: 2, DEFER: 0

#	Severity	Title	Decision
F1	MEDIUM	No canonical source for epic/story numbers	DISMISS
F2	MEDIUM	status: done not listed as recognized spec status	DISMISS

Details

F1 — DISMISS: LLM inference from epics file and slug conventions is the intended design; no canonical frontmatter field needed.

F2 — DISMISS: The "recognized value" list on line 24 is scoped to resumable statuses for routing; done is intentionally excluded because completed specs have no step to resume. Line 54 correctly uses done for a different purpose (predecessor continuity context).