wbs-verification
Skill0 starsChanged 2 months agoMITView on GitHub
Independently verify an implementation WBS for completeness, traceability, dependency correctness, source/design consistency, verification coverage, and missing work. Use before implementation readiness certification.
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
- technical-design-verificationSkill
- wbs-generationSkill
The file
Show its rules and text
We split this skill into the rules it gives an agent
13 rules. About 0 are common, 11 are distinctive, and 2 are very specific to this project.
≈0
Common · estimate
Many files like this have a rule of this kind.
≈11
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 only review tasks that exist.
- Distinctive: line 23
test design/current source
- Distinctive: line 27
Every WBS task must map to:
- Distinctive: line 45
Run this check for every WBS task, not a sample.
- Very specific: line 66
If task A references Artifact B but Artifact B owns a different primary implementation unit, mark: WBS_ARTIFACT_SEMANTIC_MISMATCH.
- Very specific: line 71
If a subordinate class/type is intended to belong to an artifact, ownership must be explicit in the registry or VERIFIED DD scope.
- Distinctive: line 97
Use current source to search likely affected surfaces implied by the verified design:
- Distinctive: line 110
Do not add a task unless the design/source evidence supports the impact.
- Distinctive: line 121
test fixture/setup prerequisites
- Distinctive: line 140
Every changed external contract and HIGH/CRITICAL behavior must have concrete verification.
- Distinctive: line 142
Do not accept unit-only coverage for externally visible changes when higher-level proof is feasible.
- Distinctive: line 175
Use the active workflow-declared WBS verification output path.
- Distinctive: line 182
If the WBS contains any implementation effort estimate such as:
≈ “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
WBS Verification
Mission
Do not only review tasks that exist. Actively search for missing implementation work.
Inputs
- authoritative requirement
- verified gap analysis
- verified TDD
- artifact registry
- verified detailed designs
- detailed-design verification reports
- WBS
- test design/current source
1. Backward Trace Every Task
Every WBS task must map to:
Requirement
→ Gap
→ TDD Decision
→ Design Artifact
→ Verified DD
→ WBS TaskTasks with no approved design basis are:
UNAPPROVED_WORK
1A. Exhaustive WBS ↔ Artifact Semantic Identity
Run this check for every WBS task, not a sample.
For each WBS row, derive the task's primary implementation deliverable from its title/objective and compare it with:
referenced Artifact ID
→ canonical registry artifact name
→ registry responsibility/scope
→ VERIFIED DD owned implementation unit(s)A PASS requires semantic identity, not just link existence.
Create:
| WBS ID | Task / Primary Deliverable | Artifact ID | Canonical Artifact Name | Registry Responsibility | DD-Owned Unit | Match? | Finding |
Hard rules:
- If task A references Artifact B but Artifact B owns a different primary implementation unit, mark:
WBS_ARTIFACT_SEMANTIC_MISMATCH. - A symbol appearing somewhere in the TDD/DD is not enough to re-label the artifact.
- The canonical artifact name/responsibility comes from the design artifact registry; do not invent a different semantic label for the same Artifact ID in another section.
- If a subordinate class/type is intended to belong to an artifact, ownership must be explicit in
the registry or VERIFIED DD scope. Mere
@see, mention, dependency, or cross-reference is not ownership. - Any HIGH/CRITICAL semantic mismatch blocks
WBS_VERIFIED.
Required threshold:
all WBS task semantic mappings valid = 100%2. Forward Coverage Every Design Artifact
For every DD_VERIFIED artifact ask:
- production/config/data changes covered?
- callers/consumers affected by signature/contract change covered?
- integration changes covered?
- migration/compatibility work covered?
- tests covered?
- docs/config samples covered if required?
Missing work is a first-class finding.
3. Source-Side Missing Work Search
Use current source to search likely affected surfaces implied by the verified design:
- callers/references
- implementations
- constructors/factories/DI
- serializers/mappers
- config consumers
- migrations/schema
- tests/fixtures
- generated/config files
- integration adapters
- sibling operations that must preserve behavior
Do not add a task unless the design/source evidence supports the impact.
4. Dependency Verification
Check:
- dependency direction
- prerequisites
- circular dependency
- sequencing of interface vs consumer changes
- migration/config sequencing
- test fixture/setup prerequisites
5. Done-Criteria Verification
Done criteria must be observable and verifiable.
Reject vague criteria like:
- "works correctly"
- "update code"
- "test as needed"
Prefer:
- build/test target
- exact contract behavior
- artifact/test IDs
- source-visible state
6. Test / Verification Coverage
Every changed external contract and HIGH/CRITICAL behavior must have concrete verification.
Do not accept unit-only coverage for externally visible changes when higher-level proof is feasible.
7. Verification Matrix
| Finding ID | WBS / Artifact | Finding | Evidence | Severity | Required Correction |
8. Gate
Before WBS_VERIFIED:
[ ] every task has approved design traceability
[ ] every task's primary deliverable semantically matches its referenced artifact responsibility / DD-owned unit
[ ] semantic identity verification was executed for 100% of WBS tasks
[ ] every verified design artifact has complete task coverage
[ ] source-side missing-work search completed
[ ] dependency DAG is implementable
[ ] every code-changing task has verification
[ ] external contract changes have appropriate higher-level tests
[ ] done criteria are concrete
[ ] no unresolved HIGH/CRITICAL WBS gap remainsOverall:
- WBS_VERIFIED
- WBS_PARTIALLY_VERIFIED
- WBS_INCOMPLETE
- WBS_BLOCKED_SOURCE_GAP
Output
Use the active workflow-declared WBS verification output path. Do not reuse legacy WBS verification artifacts.
4.0 WBS / Effort Separation Gate
WBS is not effort estimation.
If the WBS contains any implementation effort estimate such as:
- hours
- person-days
- story points
- "~1 week"
- duration estimates attached to tasks/phases
record:
WBS_CONTAINS_UNAPPROVED_EFFORT_ESTIMATE
and do NOT return WBS_VERIFIED.
Effort estimation requires a separate explicitly-invoked capability and evidence model.
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.

