Skill: Execution Alignment

Purpose

Keep project execution aligned with higher-level planning by running a post-task alignment loop: traceback from completed work to strategy, detect drift, and produce top-down recalibration recommendations.

Core Objective

Primary Goal: Produce an actionable Execution Alignment Report after task completion, with clear drift classification and prioritized recalibration actions.

Success Criteria (ALL must be met):

✅ Traceback completed: The completed task is traced through applicable project layers for the selected mode
✅ Drift classified: All detected misalignments use the drift model (goal, requirement, architecture, roadmap, priority), with impact scope and root cause per item
✅ Calibration produced: A top-down recalibration recommendation list is provided, including next tasks
✅ Report persisted: Execution Alignment Report is written to the agreed path
✅ Evidence readiness assessed: Missing or weak documentation is explicitly scored and reflected in confidence
✅ Unsafe writes avoided: If mapping is uncertain or updates are destructive, user confirmation is requested before proposing file-level changes

Acceptance Test: Can a teammate read the report and immediately understand whether execution is still aligned, what drift exists, what to do next, and why?

Scope Boundaries

This skill handles:

Post-task alignment assessment
Bottom-up traceback across project layers
Drift detection, impact analysis, and root cause labeling
Top-down recalibration suggestions
Next-task recommendations with priority hints
Evidence readiness assessment and graceful degradation when planning documents are incomplete

This skill does NOT handle:

Rewriting requirements from scratch (use analyze-requirements)
Redesigning architecture (use brainstorm-design)
Creating a new roadmap or milestone system from scratch
Implementing code changes
Team retrospective facilitation ("what went well / poorly")

Handoff point: After the report is produced and reviewed, hand off to the relevant skill by issue type (requirements, architecture, planning, or repair loop).

Use Cases

Post-task checkpoint: Validate alignment after any completed ticket
Milestone closure review: Run full alignment before marking a milestone complete
Release readiness: Detect planning drift before a release cut
Scope-shift diagnosis: Investigate whether recent work still supports current goals
Backlog reprioritization input: Generate evidence-backed next-task ordering signals

Behavior

Agent Prompt Contract

At execution start, follow this instruction contract:

You are responsible for project execution alignment.

When a task is completed, perform traceback analysis to ensure alignment
with project layers and produce a structured Execution Alignment Report.

Phase 0: Mode and Context Resolution

Resolve mode with deterministic policy:
- If user explicitly sets mode, use it
- Else if context contains release, milestone-closed, or epic-done, use Full Alignment Mode
- Else use Lightweight Mode
Resolve document paths:
- Default mapping assumes project-documentation-template layout
- Apply user path mapping overrides when provided
Confirm minimum context:
- Completed task summary
- At least one traceability anchor (requirement ID, roadmap item, milestone reference, or equivalent)

Phase 0.5: Evidence Readiness Assessment

Assess evidence quality before traceback:

strong: canonical docs exist for all required layers in selected mode
weak: partial docs exist, but at least one required layer depends on secondary evidence (issues/PR/commit notes)
missing: critical layers unavailable; traceback cannot produce reliable drift typing

Rules:

Missing docs must not silently pass as aligned.
Use secondary evidence only when canonical docs are missing, and label it explicitly.
Confidence in conclusions must be adjusted by readiness level:
- strong -> high confidence
- weak -> medium confidence
- missing -> low confidence or blocked

Mode Definitions

Lightweight Mode: Task Backlog → Roadmap → Requirements
Full Alignment Mode: Task Backlog → Roadmap → Milestones → Architecture → Requirements → Project Goals

Phase 1: Traceback (Bottom-Up)

Identify completed task intent, output, and target outcome
Walk upward through selected layers and record per-layer alignment status:
- aligned | partial | misaligned | unknown
Capture supporting evidence for each status
If critical mapping is missing, enter NeedsMappingConfirmation:
- List missing links
- Offer 1-3 candidate mappings with rationale
- Request user confirmation before continuing
- If unresolved, output blocked report with required minimum inputs
If readiness is weak or missing, annotate per-layer evidence source:
- canonical (project docs)
- secondary (issue/PR/commit context)
- none

Phase 2: Drift Detection

Classify each drift item as:

Goal Drift: Work no longer supports current project objective
Requirement Drift: Requirement changed, deprecated, or already superseded
Architecture Drift: Implementation direction diverges from current architecture decisions
Roadmap Drift: Sequencing or roadmap assumptions changed
Priority Drift: Priority is stale relative to current business direction

For each item, output:

Type
Impact Scope
Root Cause
Severity (low | medium | high)

Phase 3: Calibration (Top-Down)

Re-derive priority from top layers downward (Goals → Requirements → Architecture → Milestones → Roadmap → Backlog)
Produce recalibration actions:
- Priority adjustments
- Sequence changes
- Dependency corrections
- Follow-up analysis handoffs
Recommend next tasks with rationale and urgency window
If suggesting edits to canonical planning docs, request explicit confirmation before generating file-level change proposals

Phase 4: Persist Report

Write report to:

Default: docs/calibration/YYYY-MM-DD-task-slug-calibration.md
Or user-specified path

Report must include a machine-readable drift block (YAML or JSON) in addition to human-readable sections. Report must also include an evidence readiness block and explicit confidence level.

Input & Output

Input

Completed task description and outcome
Optional mode override (lightweight | full)
Optional document root/path mapping
Optional context: release/milestone/epic markers

Output

Execution Alignment Report Template

# Execution Alignment Report: <task title>

**Date:** YYYY-MM-DD
**Mode:** Lightweight | Full
**Status:** aligned | partial | misaligned | blocked
**Confidence:** high | medium | low

## Completed Task
- Summary:
- Outcome:

## Traceback Path
Task Backlog -> Roadmap -> Milestones -> Architecture -> Requirements -> Project Goals

## Evidence Readiness
- Readiness: strong | weak | missing
- Missing Layers:
- Secondary Sources Used:

## Alignment Status
- Goal Alignment:
- Requirement Alignment:
- Architecture Alignment:
- Milestone Alignment:
- Roadmap Alignment:

## Drift Detected
- Type:
  Impact Scope:
  Root Cause:
  Severity:

## Impact Analysis
- Delivery impact:
- Technical impact:
- Planning impact:

## Calibration Suggestions
1.
2.
3.

## Recommended Next Tasks
1.
2.
3.

## Machine-Readable Drift

    drifts:
      - driftType: "Architecture Drift"
        severity: "high"
        owner: "platform-team"
        dueWindow: "this-sprint"
        impactScope: "API gateway auth path"
        rootCause: "Service mesh ADR replaced gateway-centric pattern"
    evidence:
      readiness: "weak"
      confidence: "medium"
      missingLayers:
        - "architecture/adrs/latest-decision.md"
      secondarySources:
        - "PR#142"
        - "commit:abc1234"

Restrictions

Hard Boundaries

Do NOT invent traceability links when evidence is missing
Do NOT claim high confidence when readiness is weak or missing
Do NOT silently modify planning truth sources (goals, requirements, architecture, roadmap) without explicit user approval
Do NOT collapse drift categories into a generic "misalignment" bucket; keep typed drift outputs
Do NOT skip layer checks in Full mode unless the layer is truly unavailable (mark as unknown and explain)
Do NOT present recommendations without rationale tied to traceback evidence

Skill Boundaries (Avoid Overlap)

Do NOT do these (other skills handle them):

Requirements redefinition workflow → analyze-requirements
Architecture option design workflow → brainstorm-design
Repo-level onboarding workflow → onboard-repo
Automated test-and-fix loop execution → run-repair-loop

When to stop and hand off:

Requirements are invalid or contradictory → hand off to analyze-requirements
Architecture conflict is primary blocker → hand off to brainstorm-design
Report indicates active implementation defects requiring repair → suggest run-repair-loop

Self-Check

Core Success Criteria (ALL must be met)

Traceback completed for selected mode
Drift items typed using the five-category model
Each drift has impact scope and root cause
Top-down calibration actions provided
Report persisted to agreed path
Evidence readiness level and confidence are explicitly reported
Confirmation requested where mapping or write safety is uncertain

Process Quality Checks

Mode selection followed deterministic policy
Mapping uncertainty triggered NeedsMappingConfirmation
Canonical vs secondary evidence is clearly separated
Evidence references are present for each alignment status
Recommendations are prioritized and time-bounded
Handoffs to adjacent skills are explicit when needed

Acceptance Test

Can a teammate execute the top 1-3 recommendations without additional clarification and explain why they are highest priority?

If NO: report is incomplete; refine traceback evidence and recalibration details.

If YES: report is complete; proceed to handoff or execution planning.

Examples

Example 1: Lightweight Mode (Routine Task)

Context: Completed task updates search filter UI behavior.

Mode resolution: No explicit override, no release/milestone/epic marker → Lightweight.

Traceback:

Task aligns with roadmap item search-improvements-q2
Requirement still valid, but acceptance criteria changed last week
Evidence readiness is strong because requirement and roadmap docs are both canonical and current

Drift:

Requirement Drift (medium): Requirement wording changed to include keyboard navigation; completed task only covers click interactions

Calibration:

Add follow-up task for keyboard navigation
Move accessibility acceptance check before merging related UI tasks
Reorder backlog items to unblock regression testing

Example 2: Full Mode (Milestone Closure)

Context: Milestone tagged milestone-closed after API gateway auth rollout.

Mode resolution: Marker hit → Full.