agentleFS
tuana-vntuana-vn/system-reverse-engineer/.github/skills/technical-design-verification/SKILL.md

technical-design-verification

Skill0 starsChanged 2 months agoMITView on GitHub

Independently verify a proposed TDD against requirement, current source, verified gaps, mandatory static/dynamic architecture content, provenance/exposure evidence, compatibility, and traceability before downstream detailed design or test planning.

The file

Show its rules and text

We split this skill into the rules it gives an agent

14 rules. About 0 are common, 12 are distinctive, and 2 are very specific to this project.

  • ≈0

    Common · estimate

    Many files like this have a rule of this kind.

  • ≈12

    Distinctive · estimate

    This author's own choices. Worth a look.

  • 2

    Very specific

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

  • Common
  • Distinctive
  • Very specific
  1. Distinctive:

    Do not merely verify that the proposal sounds reasonable.

    line 11
  2. Distinctive:

    Check for the exact sections:

    line 32
  3. Distinctive:

    If structural responsibility changes, the section must contain a static Mermaid diagram.

    line 44
  4. Distinctive:

    If no static structural change exists, the section must explicitly show/describe the unchanged resulting structure and why.

    line 45
  5. Very specific:

    Do not return TDD_VERIFIED.

    line 49
  6. Distinctive:

    No design item may widen scope silently.

    line 87
  7. Distinctive:

    Before declaring the TDD verified, enumerate every implementation-significant PROPOSED unit introduced by the design.

    line 94
  8. Distinctive:

    Do not treat every helper method or local variable as a design unit.

    line 109
  9. Distinctive:

    If any implementation-significant proposed unit has no explicit owner:

    line 136
  10. Very specific:

    If a structurally material proposed unit is absent from PROPOSED Static Architecture:

    line 140
  11. Distinctive:

    When the requirement contains a MUST LOG / MUST NOT LOG decision matrix or equivalent scenario table, treat that matrix as executable …

    line 150
  12. Distinctive:

    Before the final TDD gate, verify:

    line 255
  13. Distinctive:

    no unit is "owned" only because another artifact mentions it

    line 260
  14. Distinctive:

    Use the active workflow-declared output path.

    line 293

≈ “Common” is an estimate: at least 1 in 5 of the 350,495 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

Technical Design Verification — 4.0

Mission

Do not merely verify that the proposal sounds reasonable.

Verify:

Requirement
→ Gap
→ Current Source
→ Static Architecture
→ Runtime Architecture
→ Proposed Change
→ Compatibility
→ Traceability

The verifier must NOT silently repair the TDD.


1. Mandatory Architecture Contract

Check for the exact sections:

## CURRENT Static Architecture
## PROPOSED Static Architecture
## Component Responsibilities
## PROPOSED Dynamic Architecture / Runtime Sequences

Rules:

  • PROPOSED Static Architecture is mandatory for a reviewer-grade TDD.
  • If structural responsibility changes, the section must contain a static Mermaid diagram.
  • If no static structural change exists, the section must explicitly show/describe the unchanged resulting structure and why.
  • sequence diagrams do not satisfy the static architecture requirement.

Missing applicable static architecture => TDD_PARTIALLY_VERIFIED at best. Do not return TDD_VERIFIED.


2. Diagram ↔ Source ↔ Prose Consistency

For every existing type/component in CURRENT/PROPOSED diagrams:

  • prove it exists
  • prove its relevant role/relationship

For every proposed type/component:

  • confirm it is explicitly marked PROPOSED
  • confirm responsibility is defined
  • confirm it appears in the exact change set when implementation is required

Check consistency across:

PROPOSED Static Architecture
↔ Component Responsibilities
↔ Class/Interface Design
↔ Runtime Sequences
↔ Exact Change Set

Contradictory names/relationships are verification failures.


3. Requirement / Gap Scope Verification

Create:

TDD Decision IDRequirement IDsGap IDsSurfaceProposed BehaviorIn Scope?Evidence

No design item may widen scope silently.


3A. Implementation-Significant Proposed Unit Ownership

Before declaring the TDD verified, enumerate every implementation-significant PROPOSED unit introduced by the design.

Examples include, when present:

  • new class
  • new interface
  • new data carrier / DTO / value object
  • new exception type
  • new policy / strategy / provider
  • new adapter / client
  • new configuration owner
  • new persistence object / schema
  • materially new method/contract requiring independent implementation work

Do not treat every helper method or local variable as a design unit. A unit is implementation-significant when it creates a distinct responsibility, contract, lifecycle, dependency boundary, failure behavior, configuration/state ownership, or verification obligation.

For each unit prove explicit ownership:

PROPOSED unit
→ static/component design presence when structurally material
→ explicit owning responsibility
→ intended artifact ownership:
     DEDICATED_ARTIFACT
     OR
     SUBORDINATE_TO:<future ART-ID>

Create:

Proposed UnitTypeWhy Significant?Static/Component EvidenceIntended OwnerDedicated/SubordinateOwnership Explicit?Verdict

Required threshold:

PROPOSED_UNIT_OWNERSHIP_COVERAGE = 100%

If any implementation-significant proposed unit has no explicit owner:

TDD_INCOMPLETE_PROPOSED_UNIT_OWNERSHIP

If a structurally material proposed unit is absent from PROPOSED Static Architecture:

TDD_INCOMPLETE_ARCHITECTURE

A mention, JavaDoc reference, @see, import, dependency arrow, method signature, or nearby prose does not by itself prove ownership.

3B. Authoritative Decision-Matrix Execution Gate

When the requirement contains a MUST LOG / MUST NOT LOG decision matrix or equivalent scenario table, treat that matrix as executable specification. For every TDD policy, predicate, decision tree, or branching algorithm:

  1. enumerate every authoritative scenario row;
  2. evaluate the proposed logic with representative inputs for that row;
  3. record actual proposed outcome vs required outcome;
  4. require 100% row agreement.

Create:

Requirement RowRepresentative InputsRequired OutcomeProposed Logic OutcomeMatch?Evidence

Required threshold:

DECISION_MATRIX_SEMANTIC_COVERAGE = 100%
DECISION_MATRIX_MISMATCHES = 0

A document that mentions all requirement IDs but implements a predicate that returns the wrong result for any row is not verified. Use TDD_CONTRADICTED when the design logic directly contradicts an authoritative requirement row.

3C. Requirement-Fidelity Gate

For every mandatory-looking field, behavior, scenario, or acceptance criterion in the TDD, trace it back to an authoritative requirement or label it explicitly as PROPOSED design. Source availability alone does not convert a field into a requirement.

Unsupported requirement expansion blocks TDD_VERIFIED.

4. Current Source Anchor Verification

For each existing file/class/method/config/boundary named by TDD:

  • exists?
  • current role?
  • relevant caller/callee?
  • reachable insertion point?

Unproven existing anchor:

DESIGN_SOURCE_ANCHOR_NOT_PROVEN


5. Field / Data Provenance Verification

Where data/context/output changes:

CURRENT provenance
→ proven break point
→ PROPOSED change
→ observable behavior

Reject redundant or misplaced fixes.


6. Boundary → Exposure Verification

Where integration boundary changes:

boundary
← all invocation sites
← all material callers
← all external/internal roots

No representative-endpoint shortcuts.


7. Compatibility / Exclusion Verification

Check:

  • unaffected behavior
  • MUST-NOT / non-goals
  • API/protocol behavior
  • error/result preservation
  • configuration defaults
  • internal/background flows
  • sensitive-data rules
  • failure isolation

8. Traceability Verification

RequirementGapTDD DecisionStatic ComponentRuntime FlowChangeTestVerdict

A major component/change with no backward traceability is a finding.


8A. Proposed Unit Ownership Gate

Before the final TDD gate, verify:

  • implementation-significant proposed units enumerated = 100%
  • explicit ownership assigned = 100%
  • structurally material proposed units represented in static architecture = 100%
  • no unit is "owned" only because another artifact mentions it

Any failure blocks TDD_VERIFIED.

9. Verification Gate

TDD_VERIFIED requires:

[ ] mandatory static architecture contract satisfied
[ ] diagram/prose/change-set consistency proven
[ ] every material decision maps to requirement/gap
[ ] authoritative decision matrix evaluated semantically with 100% row agreement
[ ] no unsupported requirement expansion
[ ] existing source anchors proven
[ ] proposed artifacts clearly proposed
[ ] provenance complete where applicable
[ ] exposure closure complete where applicable
[ ] compatibility/non-goals preserved
[ ] no HIGH/CRITICAL source-resolvable uncertainty

Statuses:

  • TDD_VERIFIED
  • TDD_PARTIALLY_VERIFIED
  • TDD_CONTRADICTED
  • TDD_BLOCKED_SOURCE_GAPS
  • TDD_INCOMPLETE_ARCHITECTURE
  • TDD_INCOMPLETE_PROPOSED_UNIT_OWNERSHIP

Output:

Use the active workflow-declared output path. Do not invent or reuse a legacy verification filename.

<!-- workflow-status-contract:start -->
emitted_statuses:
  - TDD_VERIFIED
  - TDD_PARTIALLY_VERIFIED
  - TDD_CONTRADICTED
  - TDD_BLOCKED_SOURCE_GAPS
  - TDD_INCOMPLETE_ARCHITECTURE
  - TDD_INCOMPLETE_PROPOSED_UNIT_OWNERSHIP
<!-- workflow-status-contract:end -->

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.