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.
In this repository33 other files it gives its agents.
- .github/copilot-instructions.mdCopilot instructions
- adversarial-baseline-auditSkill
- architecture-drift-detectionSkill
- architecture-reconstruction-and-diagramsSkill
- artifact-detailed-designSkill
- baseline-source-of-truth-maintenanceSkill
- claim-verification-and-promotionSkill
- configuration-source-traceSkill
- contract-impact-test-designSkill
- current-state-tdd-synthesisSkill
- design-artifact-decompositionSkill
- design-artifact-decomposition-verificationSkill
- detailed-design-verificationSkill
- full-reverse-engineeringSkill
- implementation-readiness-certificationSkill
- incident-root-cause-analysisSkill
- integration-selection-analysisSkill
- observability-traceability-analysisSkill
- patch-impact-reviewSkill
- persistence-schema-reverse-engineeringSkill
- post-readiness-adversarial-auditSkill
- regression-test-designSkill
- requirement-compliance-reviewSkill
- resolve-high-impact-unknownsSkill
- resume-reverse-engineeringSkill
- reverse-engineering-bootstrapSkill
- reverse-engineering-coverage-auditSkill
- run-engineering-workflowSkill
- runtime-binding-verificationSkill
- runtime-flow-analysisSkill
- technical-design-proposalSkill
- wbs-generationSkill
- wbs-verificationSkill
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
- Distinctive: line 11
Do not merely verify that the proposal sounds reasonable.
- Distinctive: line 32
Check for the exact sections:
- Distinctive: line 44
If structural responsibility changes, the section must contain a static Mermaid diagram.
- Distinctive: line 45
If no static structural change exists, the section must explicitly show/describe the unchanged resulting structure and why.
- Very specific: line 49
Do not return TDD_VERIFIED.
- Distinctive: line 87
No design item may widen scope silently.
- Distinctive: line 94
Before declaring the TDD verified, enumerate every implementation-significant PROPOSED unit introduced by the design.
- Distinctive: line 109
Do not treat every helper method or local variable as a design unit.
- Distinctive: line 136
If any implementation-significant proposed unit has no explicit owner:
- Very specific: line 140
If a structurally material proposed unit is absent from PROPOSED Static Architecture:
- Distinctive: line 150
When the requirement contains a MUST LOG / MUST NOT LOG decision matrix or equivalent scenario table, treat that matrix as executable …
- Distinctive: line 255
Before the final TDD gate, verify:
- Distinctive: line 260
no unit is "owned" only because another artifact mentions it
- Distinctive: line 293
Use the active workflow-declared output path.
≈ “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.
Files
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
→ TraceabilityThe 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 SequencesRules:
PROPOSED Static Architectureis 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 SetContradictory names/relationships are verification failures.
3. Requirement / Gap Scope Verification
Create:
| TDD Decision ID | Requirement IDs | Gap IDs | Surface | Proposed Behavior | In 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 Unit | Type | Why Significant? | Static/Component Evidence | Intended Owner | Dedicated/Subordinate | Ownership 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:
- enumerate every authoritative scenario row;
- evaluate the proposed logic with representative inputs for that row;
- record actual proposed outcome vs required outcome;
- require 100% row agreement.
Create:
| Requirement Row | Representative Inputs | Required Outcome | Proposed Logic Outcome | Match? | Evidence |
Required threshold:
DECISION_MATRIX_SEMANTIC_COVERAGE = 100%
DECISION_MATRIX_MISMATCHES = 0A 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 behaviorReject redundant or misplaced fixes.
6. Boundary → Exposure Verification
Where integration boundary changes:
boundary
← all invocation sites
← all material callers
← all external/internal rootsNo 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
| Requirement | Gap | TDD Decision | Static Component | Runtime Flow | Change | Test | Verdict |
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 uncertaintyStatuses:
TDD_VERIFIEDTDD_PARTIALLY_VERIFIEDTDD_CONTRADICTEDTDD_BLOCKED_SOURCE_GAPSTDD_INCOMPLETE_ARCHITECTURETDD_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 -->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
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.

