Agent skill

dialogue-assess-phase

Assess readiness to transition between SDLC phases. Aggregates component assessments and generates PROCEED/PROCEED_WITH_CAUTION/DEFER recommendation. Triggers on "assess phase", "phase readiness", "ready to proceed", "phase transition check", "can we move to phase".

Stars 163
Forks 31

Install this agent skill to your Project

npx add-skill https://github.com/majiayu000/claude-skill-registry/tree/main/skills/data/dialogue-assess-phase

SKILL.md

Dialogue: Phase Readiness Assessment

Assess readiness to transition from one SDLC phase to the next. This composite assessment aggregates component assessments and evaluates overall readiness, generating a recommendation that requires human approval.

Framework Grounding

This skill operationalises:

  • Phase transitions: 35-55% information loss at transitions requires explicit checkpoints
  • Theory-building: Validates sufficient understanding exists before proceeding
  • STS joint optimisation: Human approval ensures both technical and social readiness

When to Use

Use this skill at phase transition points:

  • Phase 1 (Initiation/Conception) → Phase 2 (Planning)
  • Phase 2 (Planning) → Phase 3 (Analysis/Requirements)
  • Phase 3 (Analysis/Requirements) → Phase 4 (Design/Architecture)
  • And so on through Phase 7 (Deployment/Operations)

Phase Readiness Dimensions

The assessment evaluates four readiness dimensions:

Dimension What It Measures Key Inputs
documentation_readiness Are phase artefacts complete? Required documents exist
knowledge_transfer_readiness Is knowledge documented/shared? Theory captured, decisions logged
stakeholder_readiness Are stakeholders aligned? Alignment assessment, approvals
technical_readiness Are technical prerequisites met? Tests pass, dependencies resolved

Recommendation Outcomes

Recommendation Meaning Action
PROCEED All dimensions satisfactory Human approves, transition proceeds
PROCEED_WITH_CAUTION Minor gaps identified Human reviews gaps, may proceed with mitigations
DEFER Significant gaps present Address blockers before transitioning

How to Assess Phase Readiness

Step 1: Gather Context

First, gather information about the current state:

bash
# Check for recent problem framing assessment
ls -la ${CLAUDE_PROJECT_DIR}/.dialogue/logs/assessments/ASSESS-*.yaml | tail -5

# Check for recent daily checks
grep -l "assessment_type: DAILY_CHECK" ${CLAUDE_PROJECT_DIR}/.dialogue/logs/assessments/*.yaml | tail -5

# Check decision log activity
ls -la ${CLAUDE_PROJECT_DIR}/.dialogue/logs/decisions/ | tail -10

Step 2: Interactive Assessment

Ask the user to evaluate each dimension using AskUserQuestion:

  1. Documentation readiness (1-5): Are required phase artefacts complete?
  2. Knowledge transfer readiness (1-5): Is knowledge documented and shared?
  3. Stakeholder readiness (1-5): Are stakeholders aligned on proceeding?
  4. Technical readiness (1-5): Are technical prerequisites met?

Then ask: 5. What is the current phase? (1-7) 6. What is the target phase? (2-7) 7. Are there any blockers? (optional free text) 8. Are there any risks to proceeding? (optional free text)

Step 3: Reference Component Assessments

If available, reference recent component assessments:

  • Problem framing assessment ID (if exists)
  • Stakeholder alignment assessment ID (if exists, from FW-041)
  • TTKM assessment ID (if exists, from FW-041)
  • Recent daily check IDs

Step 4: Log the Assessment

bash
${CLAUDE_PLUGIN_ROOT}/skills/dialogue-assess-phase/scripts/log-assess-phase.sh \
  <assessor> \
  <current_phase> <target_phase> \
  <documentation_readiness> <knowledge_transfer_readiness> \
  <stakeholder_readiness> <technical_readiness> \
  [problem_framing_ref] [blockers] [risks]

Parameters

Parameter Values Description
assessor ai:claude or human:<id> Who performed the assessment
current_phase 1-7 Current SDLC phase
target_phase 2-7 Target phase (must be > current)
documentation_readiness 1-5 Documentation completeness
knowledge_transfer_readiness 1-5 Knowledge sharing quality
stakeholder_readiness 1-5 Stakeholder alignment
technical_readiness 1-5 Technical prerequisites
problem_framing_ref string (optional) ASSESS-... ID of framing assessment
blockers string (optional) Blocking issues (comma-separated)
risks string (optional) Identified risks (comma-separated)
context string (optional) Situational context
tags string (optional) Comma-separated categorisation tags

Recommendation Logic

The script computes a recommendation based on dimension scores:

Average score = (doc + knowledge + stakeholder + technical) / 4

If average >= 4.0 AND no blockers AND problem_framing exists:
  → PROCEED
Else if average >= 3.0 AND blockers are manageable:
  → PROCEED_WITH_CAUTION
Else:
  → DEFER

Special conditions:

  • Missing problem framing assessment: Maximum recommendation is PROCEED_WITH_CAUTION
  • Any dimension score of 1: Forces DEFER
  • Blockers present: Maximum recommendation is PROCEED_WITH_CAUTION

DEFER Remediation Guidance

When the recommendation is DEFER, the assessment includes a defer_guidance block that provides actionable remediation advice based on two factors:

Factor 1: Phase Tacit Percentage

The current phase's information composition determines remediation character:

Phase Tacit % Remediation Character
1. Initiation 75% Primarily dialogue-based
2. Planning 55% Dialogue-dominant
3. Requirements 50% Balanced
4. Design 40% Balanced, slightly artifact-weighted
5. Implementation 35% Artifact-dominant
6-7. Testing/Ops 30% Primarily artifact/process

Factor 2: Gap Dimension

The lowest-scoring dimension identifies what type of gap to address:

Gap Meaning Natural Remediation
documentation Artifacts incomplete Review/create documents
knowledge Tacit understanding not shared Dialogue, pairing, mentoring
stakeholder Alignment issues Facilitation, workshops
technical Blockers, dependencies Technical resolution

Recommended Approach

The combination produces a recommended_approach:

  • DIALOGUE: High-tacit phase or knowledge/stakeholder gaps → convene sessions, elicit understanding
  • MIXED: Medium-tacit phase or mixed gaps → balance dialogue with artifact work
  • ARTIFACT: Low-tacit phase with documentation/technical gaps → focus on specifications, resolution

Example DEFER Output

yaml
defer_guidance:
  primary_gap: knowledge
  primary_gap_score: 2
  secondary_gaps: []
  phase_tacit_percentage: 75
  recommended_approach: DIALOGUE
  specific_actions:
    - "Conduct knowledge transfer sessions with domain experts"
    - "Pair with experienced team members"
    - "Facilitate stakeholder interviews to capture tacit understanding"
  restart_point:
    phase: 1
    phase_name: "Initiation/Conception"
    focus: "Address knowledge gaps before re-assessing"

Human Approval Workflow

Phase transitions require human approval:

  1. AI presents assessment summary with recommendation
  2. Human reviews:
    • Component assessment references
    • Dimension scores and rationale
    • Identified blockers and risks
  3. Human decides: approve, approve with conditions, or defer
  4. Decision is logged with approval metadata

Example Approval Flow

AI: Phase readiness assessment complete.

    Current Phase: 1 (Problem Definition)
    Target Phase:  2 (Requirements)

    Dimension Scores:
    - Documentation:       4/5 (requirements doc drafted)
    - Knowledge Transfer:  4/5 (decisions logged, theory captured)
    - Stakeholder:         3/5 (one stakeholder pending review)
    - Technical:           5/5 (no blockers)

    Component Assessments:
    - Problem Framing: ASSESS-20260122-140000 (5/6, confidence 4)

    Blockers: None
    Risks: Stakeholder review may surface new requirements

    Recommendation: PROCEED_WITH_CAUTION

    Do you approve proceeding to Phase 2?
    [Options: Approve / Approve with conditions / Defer]

User: Approve with conditions - complete stakeholder review within 48 hours

AI: Approved with conditions. Logging the decision...
    [Logs DEC-... with approval and conditions]

    Phase transition approved. Moving to Phase 2 (Planning).
    Condition: Complete stakeholder review within 48 hours.

Output

The script returns the generated assessment ID (e.g., ASSESS-20260122-150000).

The assessment is stored in .dialogue/logs/assessments/ and creates:

  • Assessment YAML file with full readiness data
  • Context graph node (ARTIFACT with artifact_type: ASSESSMENT)
  • CREATED edge from assessor to assessment
  • ASSESSES edges to referenced component assessments

Example Direct Invocation

bash
${CLAUDE_PLUGIN_ROOT}/skills/dialogue-assess-phase/scripts/log-assess-phase.sh \
  "human:pidster" \
  1 2 \
  4 4 3 5 \
  "ASSESS-20260122-140000" \
  "" \
  "Stakeholder review may surface new requirements"

Phase-Specific Considerations

Phase 1 → 2 (Initiation/Conception → Planning)

Critical inputs:

  • Problem framing assessment (strongly encouraged)
  • Stakeholder alignment on business case
  • Strategic rationale documented

Phase 2 → 3 (Planning → Analysis/Requirements)

Critical inputs:

  • Project plan and resource allocations
  • Risk register established
  • Governance structure defined

Phase 3 → 4 (Analysis/Requirements → Design/Architecture)

Critical inputs:

  • Requirements documented and reviewed
  • Stakeholders approved requirements
  • Key technical constraints identified

Phase 4 → 5 (Design/Architecture → Implementation/Construction)

Critical inputs:

  • Architecture decisions documented (ADRs)
  • Technical feasibility validated
  • Performance/scalability requirements addressed

Later Phases

Similar patterns—each transition validates that:

  1. Phase artefacts are complete
  2. Knowledge is captured and shared
  3. Stakeholders are aligned
  4. Technical prerequisites are met

Schema Reference

See Assessment Schema for the complete PHASE_READINESS response schema and validation rules.

Expand your agent's capabilities with these related and highly-rated skills.

Didn't find tool you were looking for?

Be as detailed as possible for better results