Recipe Update Doc

Explicit User Instruction: The user explicitly instructs and authorizes every subagent call named in this recipe. Execute each applicable call when its prerequisites are met.

Execute Skill: llm-friendly-context before writing Agent prompts, handoffs, or generated artifacts. Execute Skill: subagents-orchestration-guide before making workflow decisions, invoking agents, or resolving findings.

Context: Dedicated to updating existing design documents.

Orchestrator Definition

Core Identity: "I am an orchestrator." (see subagents-orchestration-guide skill)

Local authority gate: Make this recipe's workflow decisions and validate each returned result directly; delegate semantic deliverable production to the named specialist.

Review Resolution Gate [MANDATORY]: Resolve every actionable deliverable-review finding through subagents-orchestration-guide Review Resolution before correction or progression. Before the first finding disposition, read references/review-resolution.md from the loaded subagents-orchestration-guide skill.

Execution Gate: Complete Steps 1-6 in order, following only the branches activated by document type and review result. Advance only through each step's stated evidence, review convergence, or approval condition. Complete after the final approval gate and every applicable Completion Criterion is satisfied.

Execution Protocol:

1. Invoke named specialists for deliverable production — pass deliverable paths between them and validate their results (see subagents-orchestration-guide "Orchestrator Execution Boundary")
2. Execute update flow:
   • Identify target → Clarify changes → Update document → Review → Consistency check
   • Stop at the [Stop: Final approval] marker → Wait for user approval before completing
3. Scope: Complete when the user approves completing the update

At each Agent invocation below, build the prompt as a mechanical extraction: copy the named source values into the exact fields, apply only the declared serialization, then invoke immediately.

CRITICAL: Execute document-reviewer — it is the quality gate for document accuracy.

Workflow Overview

Copy code
Target document → change clarification
                        ↓
              technical-designer / technical-designer-frontend / prd-creator (update mode)
                        ↓ (Design Doc only)
              code-verifier → document-reviewer
                        ↓ (Design Doc only)
              design-sync → [Stop: Final approval]

Scope Boundaries

Included in this skill:

• Existing document identification and selection
• Change content clarification
• Document update with appropriate agent (update mode)
• Document review with document-reviewer
• Consistency verification with design-sync (Design Doc only)

Out of scope (redirect to appropriate skills):

• New requirement analysis
• Work planning or implementation

Responsibility Boundary: This skill completes when the user approves completing the update.

Target document: $ARGUMENTS

Execution Flow

Step 1: Target Document Identification

Discover the existing documents under docs/design/, docs/prd/, and docs/adr/.

Decision flow:

Step 2: Document Type and Layer Determination

Determine type from document path, then determine the layer to select the correct update agent:

Layer detection (for Design Doc and ADR): Read the document and determine its layer from content signals:

• Frontend (→ technical-designer-frontend): Document title/scope mentions React, components, UI, frontend; or file contains component hierarchy, state management, UI interactions
• Backend (→ technical-designer): All other cases (API, data layer, business logic, infrastructure)

ADR Update Guidance:

• Minor changes (clarification, typo fix, small scope adjustment): Update the existing ADR file
• Major changes (decision reversal, significant scope change): Create a new ADR that supersedes the original

Step 3: Change Content Clarification

Determine which sections need updating, the reason for the change, and the expected outcome after the update. Derive them from the request and the target document. Use AskUserQuestion only for an item the request and document leave undetermined, and only when a different answer would change which sections are updated or what the update must achieve.

Pass the resulting items, covered sections, and any stated total size budget to update or revision agents. Before the next approval gate, map every diff hunk to an approved item or required consistency update. Remove accidental unmapped changes; when a necessary change would alter the requested document outcome or explicit size constraint, return to the request clarification gate.

Step 4: Document Update

Invoke the update agent determined in Step 2:

Copy code
subagent_type: [Update Agent from Step 2]
description: "Update [Type from Step 2]"
prompt: |
  Operation Mode: update
  Existing Document: [path from Step 1]

  ## Changes Required
  [Step 3 statements for the sections to update, reason, and expected outcome, copied verbatim]

  Update the document to reflect the specified changes.
  Add change history entry.

Step 5: Document Review

For Design Doc updates only: Before document-reviewer, invoke code-verifier:

Copy code
subagent_type: code-verifier
description: "Verify updated Design Doc"
prompt: |
  doc_type: design-doc
  document_path: [path from Step 1]
  Verify the updated Design Doc against current codebase.

  Verification focus: Pay special attention to literal identifier referential
  integrity in the updated sections (paths, endpoints, type names, config keys).

Store output as: $CODE_VERIFICATION_OUTPUT

Invoke document-reviewer with the applicable exact shape:

• Design Doc: doc_type: DesignDoc, target: [path from Step 1], review_context: update, and verification_evidence: $CODE_VERIFICATION_OUTPUT.
• PRD: doc_type: PRD and target: [path from Step 1].
• ADR: doc_type: ADRBatch, targets: [path from Step 1], and review_context: update.

For each type, review consistency of the changed sections and their dependent statements, governing requirements, and change history.

Store output as: $STEP_5_OUTPUT

On review result:

• Pass → Proceed to Step 6
• Needs revision → Apply the Review Resolution Gate. Return to Step 4 when apply findings exist, using the following prompt:

  Copy code
  subagent_type: [Update Agent from Step 2]
  description: "Revise [Type from Step 2]"
  prompt: |
    Operation Mode: update
    Existing Document: [path from Step 1]
  
    ## Adjudicated Review Findings
    [complete reviewer finding objects verbatim, with only their orchestrator dispositions added]
  
    Treat these findings as the complete revision scope and preserve adjacent content.
  

• On re-review pass prior_feedback as [{id, disposition, reason?, evidence}]
• All actionable findings are decline → Proceed to Step 6

Step 6: Consistency Verification and Final Approval [Stop: Final approval]

For Design Doc only, invoke design-sync first:

Copy code
subagent_type: design-sync
description: "Verify consistency"
prompt: "source_design: [path from Step 1]"

When conflicts are detected, apply the Review Resolution Gate and follow its bounded verifier handoff and convergence rules, returning selected corrections to Step 4 for the owning document.

For every document type, present the updated document, the review outcome, any resolved declines, and the sync result when one ran. This is the only approval gate in the flow: wait for the user's decision before completing.

Error Handling

Completion Criteria

• Identified target document
• Change content determined from the request, the document, or a question whose answer changed the update
• Updated document with appropriate agent (update mode)
• Executed code-verifier before document-reviewer (Design Doc only)
• Executed document-reviewer and addressed feedback
• Executed design-sync for consistency verification (Design Doc only)
• Obtained user approval to complete the update

Output Example

Document update completed.

• Updated document: docs/design/[document-name].md
• Approval: user approved completing the update

When findings were declined during review, append their IDs, governing reasons, and evidence to this completion response.