agentleFS
tikalktikalk/adlc-team-skills/skills/product/product-clarify/SKILL.md

product-clarify

Skill145 starsChanged 23 days agoMITView on GitHub

Use when PDRs need refinement, conflict detection, or approval to Accepted status before /product-implement.

The file

We split this skill into the rules it gives an agent

12 rules. About 0 are common, 7 are distinctive, and 5 are very specific to this project.

  • ≈0

    Common · estimate

    Many files like this have a rule of this kind.

  • ≈7

    Distinctive · estimate

    This author's own choices. Worth a look.

  • 5

    Very specific

    Tied to their project's names, files or tools. Probably skip.

  • Common
  • Distinctive
  • Very specific
  1. Very specific:

    After /product-specify or /product-init to refine initial PDRs

    line 19
  2. Very specific:

    Before /product-implement to approve PDRs (required — implement skips non-Accepted)

    line 20
  3. Very specific:

    No PDRs exist (use /product-specify or /product-init first)

    line 26
  4. Very specific:

    Check if the draft being reviewed uses the lightweight draft template (indicated by presence of type, evidence, source, revisit-when fields …

    line 50
  5. Distinctive:

    Read the draft's captured fields (Context, Decision, Rejected Alternatives, Reason)

    line 56
  6. Distinctive:

    If already full format, proceed with normal review.

    line 66
  7. Distinctive:

    Check each PDR against standards:

    line 70
  8. Distinctive:

    HIGH: No alternatives, missing consequences

    line 113
  9. Very specific:

    Write every clarification in Simplified Technical English (short sentences, one idea each, active voice, imperative asks, named referents, …

    line 119
  10. Distinctive:

    Summarize changes: after each answer

    line 147
  11. Distinctive:

    Preserve original intent: when updating wording

    line 235
  12. Distinctive:

    Ask one question at a time

    line 240

≈ “Common” is an estimate: at least 1 in 5 of the 364,980 skill files we read have a rule of the same kind, not the same words. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.

Keep what fits and drop the rest. Your version, ready for your agents.
Files
SKILL.md

product-clarify

What this skill does

Reviews existing PDRs for quality gaps, asks targeted clarification questions, and promotes approved PDRs to Accepted status. This is the quality gate before PRD generation.

Input: Individual PDR-*.md files in .adlc/drafts/pdr/ (status Proposed or Discovered)

Output: Updated PDR-*.md files (status Accepted where approved), regenerated pdr.md index

When to use

  • After /product-specify or /product-init to refine initial PDRs
  • Before /product-implement to approve PDRs (required — implement skips non-Accepted)
  • Periodic PDR review before milestones
  • Resolving inconsistency flags from cross-feature-area analysis

When NOT to use

  • No PDRs exist (use /product-specify or /product-init first)

Execution Steps

Phase 1: Load PDRs

  1. Run setup script:
sh: scripts/bash/setup-product-clarify.sh [--json]
ps: scripts/powershell/setup-product-clarify.ps1
  1. Read all PDR files from {REPO_ROOT}/.adlc/drafts/pdr/PDR-*.md
  2. Read constitution from {REPO_ROOT}/docs/adlc/memory/constitution.md, falling back to legacy {REPO_ROOT}/.adlc/memory/constitution.md (dual-read), if either exists
  3. Build inventory:
| PDR | Title | Status | Category |
|-----|-------|--------|----------|
| PDR-001 | Target Market | Proposed | Problem |
| PDR-002 | Primary Persona | Proposed | Persona |

Step 0: Detect Lightweight Draft Format

Check if the draft being reviewed uses the lightweight draft template (indicated by presence of type, evidence, source, revisit-when fields in frontmatter and ## Rejected Alternatives / ## Reason body sections without the full formal template sections).

If lightweight:

  1. Read the draft's captured fields (Context, Decision, Rejected Alternatives, Reason)
  2. Transform to full formal template:
    • ADR: full MADR format with Decision Drivers, Considered Options, Pros/Cons, Constitution Alignment, Related ADRs
    • PDR: full PDR format with Market Forces, Consequences, Alternatives Considered, Links
    • ChDR: full ChDR format with Issue Links, Commits, Consequences, Evidence
    • CDR: full CDR format with Context Type, Target Module, Descriptor, Evidence
    • EVAL: full eval format with Error Analysis, Pass/Fail Examples, Implementation Notes
  3. Enrich from session context (add details the lightweight draft may have omitted)
  4. Present the enriched draft for review

If already full format, proceed with normal review.

Phase 2: PDR Quality Analysis

Check each PDR against standards:

DimensionCheckSeverity if Missing
ContextProblem clearly statedMEDIUM
DecisionActionable, testableHIGH
ConsequencesPositive AND negativeHIGH
Success MetricsDefined with targetsHIGH
AlternativesAt least 2 with neutral trade-offsHIGH
ConstitutionAligns with visionCRITICAL

Quality checklist:

  • Clear context explaining the problem/opportunity
  • Explicit, actionable decision statement
  • Positive AND negative consequences documented
  • At least 2 alternatives with neutral trade-offs (not "rejected because")
  • Success metrics defined
  • Risks identified with mitigation strategies
  • Valid status value
  • No conflicts with other PDRs
  • Alignment with constitution/vision principles

Phase 3: Cross-PDR Consistency

  1. Conflicting Decisions: Same concept, different decisions across PDRs
  2. Missing Dependencies: PDRs that should reference each other
  3. Terminology Drift: Same concept named differently
  4. Priority Conflicts: One PDR says B2B, another assumes B2C

Phase 4: Gap Identification

Generate gap report:

## PDR Clarification Report

| PDR | Title | Gap Type | Severity |
|-----|-------|----------|----------|
| PDR-001 | [Title] | Missing alternatives | HIGH |
| PDR-002 | [Title] | Incomplete consequences | MEDIUM |

Severity:

  • CRITICAL: Constitution violations, missing decision statement, no success metrics
  • HIGH: No alternatives, missing consequences
  • MEDIUM: Incomplete risks, unclear context
  • LOW: Minor phrasing

Phase 5: Interactive Refinement

Write every clarification in Simplified Technical English (short sentences, one idea each, active voice, imperative asks, named referents, plain words). Read the contract at skills/authoring/references/ste-rules.md before drafting questions.

Count crossed boundaries (record, lane, team, now) for each gap. See the tier rule in the shared reference:

  • 0–1 boundaries: present one clarification at a time in plain STE prose, using this format, however severe the gap:
## Clarification [N]: PDR-XXX — [Gap Type]

**Context**: [One STE sentence quoting the PDR content]

**Gap**: [One STE sentence describing the missing element]

**Question**: [One direct imperative STE question]

**Suggested Options**:
- **A**: [Option text]
- **B**: [Option text]
- **C**: [Custom response]
  • 2 or more boundaries: draw the full ASCII decision map from the reference (WHAT/WHY, crossed boundaries, reach fan-out, candidates × discriminating axes, IF-UNDECIDED). Render unknown gaps as ? slots, never prose complaints.

Rules:

  • Present one clarification at a time
  • Prioritize by severity — CRITICAL/HIGH first (severity orders questions; it never upgrades a 0–1-boundary item to a map)
  • Limit to 5 clarifications per session
  • Allow user to skip non-critical clarifications
  • Summarize changes after each answer

Phase 6: Constitution Violations (if any)

For Duplicates:

## Constitution Duplication Detected

**PDR**: PDR-XXX — [Title]
**Constitution Principle**: §[Section] — [Principle]

**Options**:
| Option | Action |
|--------|--------|
| A | Remove PDR |
| B | Convert to Reference |
| C | Add Context |
| D | Extend |

For Violations (⭐ RECOMMENDED: Amend Constitution):

## Constitution Violation Detected ⭐

**⭐ RECOMMENDED: A. Amend Constitution**
Update constitution §[Section] to accommodate this decision.

**Alternatives**:
B. Override in PDR — Document justification
C. Revise PDR — Change decision to comply
D. Remove PDR — Delete and follow constitution

Phase 7: Update PDR Files

  1. Apply clarifications per-file
  2. Update status where applicable
  3. Update "Last Updated" timestamps
  4. Regenerate pdr/pdr.md index — call generate_pdr_index drafts from pdr-lib.sh (or Generate-PdrIndex drafts from pdr-lib.ps1). This uses a frontmatter-primary + heading-fallback parser that handles both H2 and H3 metadata, preventing the blank-row bug.

Phase 8: PDR Approval ⭐

Critical: /product-implement only processes PDRs with Accepted status.

## PDR Approval ⭐

**Total PDRs**: [N]
**Status Distribution**:
- Accepted: [N]
- Proposed: [N]
- Discovered: [N]

**Options**:
| Option | Action |
|--------|--------|
| A | Accept All — Change Proposed/Discovered → Accepted |
| B | Review Specific — Select individual PDRs |
| C | Defer — Keep current status |

Bulk approval (Option A):

## Confirm Bulk Approval

Change [N] PDRs to Accepted?

| PDR | Current | New |
|-----|---------|-----|
| PDR-001 | Proposed | Accepted |
| PDR-002 | Discovered | Accepted |

Post-approval:

## PDRs Approved

**Ready for Implementation**:
- Accepted PDRs: [N]
- Pending: [N]

Run `/product-implement` to generate PRD.md.

Key Rules

Non-Destructive Refinement

  • Never delete existing PDRs without explicit user approval
  • Preserve original intent when updating wording
  • Add, don't replace consequences and alternatives
  • Mark changes with updated timestamps

Focused Clarification

  • Ask one question at a time
  • Make questions specific and answerable
  • Provide suggested options when possible
  • Respect user's time

Constitution Authority

  • Constitution violations are always flagged
  • PDRs cannot override MUST principles without justification
  • Suggest constitution updates if conflict is systemic

Configuration

  • REPO_ROOT — resolved by the setup script (searches upward for .adlc); always use {REPO_ROOT}-anchored paths, never bare relative paths. When running helpers from zsh, source the scripts/bash/ file under bash -c or emulate -L sh: the scripts rely on BASH_SOURCE[0] for path resolution and will not locate helpers under zsh.
  • PDR_DRAFTS_DIR — {REPO_ROOT}/.adlc/drafts/pdr
  • PDR_INDEX — {REPO_ROOT}/.adlc/drafts/pdr/pdr.md
  • CONSTITUTION — {REPO_ROOT}/docs/adlc/memory/constitution.md (legacy .adlc/memory/constitution.md fallback)
  • move_pdr/next_record_number accept bare IDs (8), zero-padded IDs (008), and full suffixed stems (008-amendment-2): the PDR- prefix is optional and callers must never pre-strip it.

12-Factor Alignment

  • Factor III (Mission Definition): Ensures product mission is well-defined before execution
  • Factor VII (Quality Gates): Applies human judgment before automated generation

Common Rationalizations

RationalizationReality
"The PDRs look fine — let's skip clarify."Implement will skip non-Accepted PDRs, silently dropping decisions from the PRD.
"I'll just accept all at once."Bulk accept without review risks embedding bad decisions into the PRD.
"A warning on constitution is basically okay."Constitution violations are CRITICAL because they undermine product coherence.

Red Flags

  • Skipping the approval step — without Accepted status, implement ignores the PDR.
  • Accepting without reading — defeats the purpose of the quality gate.
  • Not flagging cross-PDR conflicts — inconsistent PDRs produce an incoherent PRD.

Verification

  • All PDR files read and inventoried
  • Quality checklist applied to each PDR
  • Cross-PDR consistency verified
  • Constitution alignment checked
  • Gaps identified and prioritized
  • Clarifications applied per-file
  • pdr.md index regenerated
  • Approval phase completed (Accept/Review/Defer)
  • Accepted PDRs have status "Accepted" in their files
  • No duplicate or missing PDR IDs

Download now (ZIP)

Does it work?

Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.

No reviews yet

People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.

  • It worked
  • Partly
  • It didn't
What did you use it for? e.g. Reviewing pull requests in a Rails app

Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.

Your trail

Files you open will collect here, so you can find them again.

Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.

Now do this for your own.

Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.